Vadym Volkov Magnetron Based Radar Systems for …cdn.intechweb.org/pdfs/10315.pdf · Magnetron Based Radar Systems for Millimeter Wavelength Band – Modern Approaches and Prospects

Magnetron Based Radar Systems for Millimeter Wavelength Band – Modern Approaches and Prospects 459

Magnetron Based Radar Systems for Millimeter Wavelength Band – Modern Approaches and Prospects

Vadym Volkov

x

Magnetron Based Radar Systems for Millimeter Wavelength Band – Modern

Approaches and Prospects

Vadym Volkov Institute of Radio Astronomy of National Academy of Sciences of Ukraine

Ukraine

In honor of my mother, Ninel.

1. Introduction

Historically magnetrons were one of the first devices used to build radar systems. Namely, successful development and utilization of the magnetrons in radars had assisted essentially Allies to win battle for air and see during World War II (Brown, 1999). And definitely they were the first devices, which have allowed developing the radars operating within millimeter wavelengths region. It had happened due to these microwave sources are characterized by a number of advantages. They are rather simple in production. The magnetrons provide both high peak and average power at relatively low operational voltages as well a fair frequency potential. Because of the above advances, the magnetron remains the most utilized type of microwave vacuum tubes until now. Namelly, by virtue of the utilization magnetrons in ovens, millions people over whole world have learned the word “microwave”! However by the middle of 60th magnetron based radar systems had ceased to meet increasing performance requirements and their development had been curtailed. They had moved considerably to a niche of simple, low cost radars for great demand applications like that for marine navigation. It had happened due to both fundamental peculiarities of the magnetron operation and issues concerning their manufacturing. At first, the magnetron is an oscillator providing no modulation capabilities except the simplest case of pulse modulation. It results in difficulties to introduce advanced signal processing into the magnetron based radars. Next, uncertainty in the conceptual ability of the magnetron to produce oscillations with appropriate short –term frequency stability; considerable difficulties to develop the corresponding highly stable modulators; as well as an analog implementation of a coherence-on-receiver technique did not allow to fetch an actual potential of magnetron based system out. It resulted in the strong opinion that the magnetrons are generally not suitable to build radar systems with relevant Doppler capabilities. Further, a high spatial resolution of magnetron based systems can be achieved practically only by reducing the duration of radiated RF pulse. However, it leads to the

23

www.intechopen.com

Microwave and Millimeter Wave Technologies: Modern UWB antennas and equipment 460

following difficulties: (i) the magnetron is a highly resonant device, which limits a minimal possible pulse duration; and (ii) in order to keep the radar potential, pulse repetition rate should be high enough respectively, which reduces unambiguous radar range or requires utilization of dedicated technique to resolve range ambiguity. Further, the first millimeter wavelengths magnetrons demonstrated a low reliability making the result of their utilization rather discouraging in the most cases. Thus, appearance by the middle of 60th efficient power amplifiers based on both vacuum tubes and solid-state devices and great expectations for a rapid progress in their development as well as the introduction a pulse compression technique had given up the magnetrons for lost to use in the high performance radars. However, since recent time, magnetrons are considered again as a rather attractive choice to develop systems for millimeter wavelengths band namely. This turn has become possible due to: (i) a lack or low availability of other power devices operating within the indicated frequency range; (ii) a significant improvement in magnetrons characteristics, partially, incredible increase in their lifetime; (iii) a dramatic progress in digital signal processing technique; (iv) achievements in the development of high voltage modulators and millimeter wavelengths technique; and (v) a strong demand for millimeter wavelengths radars from non-military applications, which means a great interest in cost effective solutions. Despite a rather simple internal structure, the magnetron is characterized by a great complexity of the processes taken place inside it. There is no a more or less comprehensive theory of magnetrons until now. Numerical simulation can be accounted as very superficial. In general we prefer to treat the magnetron as something like a magic box characterized often by unpredictable and even surprising behavior. Thus, generally development of modern magnetron based radar, especially operating within millimeter wavelength region, requires profound understanding of principles of the magnetron operation, a great experience, and the utilization of specific design approaches, at a system level partially. However, often the magnetron is considered as old, well known device and respectively radars based on it are designed in a humble way. It results in a humble performance certainly. Probably, a rather uptight attitude to utilize the magnetrons to build contemporary high performance radars is caused by the above reason. In this paper we do not try to review comprehensively the current state of affairs or to cover as much as possible wide range of issues concerning the development of both millimeter wavelengths magnetrons and radars based on them. Instead, relying on own experience we draw attention to the fact that some noted disadvantages preventing the magnetron utilization in the high performance radar systems are essentially weakened until now, whereas others can be managed successfully by gaining from the achievements of modern electronics. The corresponding design approaches, which have assisted us to develop successfully a number of contemporary magnetron based radar systems operating within millimeter wavelength band and addressed to different application areas, will be disclosed more or less systematically. We hope that the bellow consideration will be helpful for radar designers to keep always in mind the possibilities providing by good old magnetron!

2. Magnetrons in radars – brief overview

We would not like to discuss here the physical principles of magnetron operation. They can be found in a variety of manuscripts (Okress, 1961;Tsimring, 2007). For us it is important

that the magnetron is a vacuum crossed field tube, which is capable to produce high power microwave oscillations with a high efficiency, and hence can be adopted conceptually to use in radar transmitter. As fairly noticed in (Skolnik, 2008), a choice of electronic device for the transmitter end stage defines practically completely radar structure and design approaches. Thus, let us outline the most important peculiarities of the magnetrons as related to their utilization in the radars. At first, the magnetrons are characterized fundamentally by a high both peak and average power. Typical values exceed 100 kW and 100 W for Ka band and 4 kW and 4 W for W and G bands correspondingly. It allows utilizing constant frequency pulse, being the simplest possible among radar signals, while keeping an appropriate radar sensitivity without the usage of a sophisticated signal processing technique. Next, any magnetron is an oscillator rather than an amplifier. It means partially that its output signal depends only on the physical layout of the magnetron internals and the parameters of its circumstance, i.e. applied voltage, a strength of magnetic fields, a value of voltage standing wave ratio at the output flange etc. In addition it is practically impossible to manipulate either parameter of the magnetron output signal independently from the other. Further, the magnetron is characterized by a highly resonant design basically. The magnetron oscillations frequency is essentially defined by electromagnetic properties of its internal layout and can be varied within a wide enough range only by changing a mechanical configuration of such layout, i.e. slowly. All above constrict evidently modulation capabilities providing by the magnetron. Three types of modulation are used in modern radar systems commonly – pulse (as a particular type of amplitude); frequency; and phase modulation respectively. Practically it may be considered that the magnetron by itself provides no ability for a fast, highly reproducible, and well-controlled phase/frequency modulation and adopts only the simplest pulse modulation. Probably only a bandwidth of electrical frequency chirp provided by W and G band magnetrons may appeal to use in the high resolution radars (see Section 0). Notice, since the magnetron output pulse is shaped at a radio frequency directly, it occupies it twice wider frequency band than it is required to ensure a definite spatial resolution. As for any other oscillator the magnetron oscillation frequency is subjected to fluctuations. According to common approach, fast and slow fluctuations are considered separately and referred as phase noise and frequency stability respectively. Concerning to radar performance the first defines quality of Doppler processing whereas the second is not so important generally except a number of rather special cases. Certainly, the total frequency variation range should not be too big as well as matching between transmitter and receiver frequencies should be ensured ever. Usually the related magnetron performance is characterized as poor and this fact is a byword to utilize such devices in the radars. On other hand the maximum possible variation of magnetron operational frequency including manufacturing tolerances is less than ± 1% over all millimeter wavelength bands in use even for non-tunable devices and cannot be considered as a serious issue. As for the phase noise, which is referred as a pulse-to-pulse frequency instability in the magnetron based systems, reflecting peculiarities of Doppler processor implementation, the following should be mentioned. Prima facie, it seems to be difficult actually expecting a high pulse-to-pulse frequency stability for the magnetrons. Q-factor of the magnetron resonant system is relatively low even for coaxial devices especially within millimeter wavelengths band as well as frequency pushing is rather big due to a high electron density inside an interaction space. Generally, the behavior of electron cloud has a considerable noise component and not

www.intechopen.com


following difficulties: (i) the magnetron is a highly resonant device, which limits a minimal possible pulse duration; and (ii) in order to keep the radar potential, pulse repetition rate should be high enough respectively, which reduces unambiguous radar range or requires utilization of dedicated technique to resolve range ambiguity. Further, the first millimeter wavelengths magnetrons demonstrated a low reliability making the result of their utilization rather discouraging in the most cases. Thus, appearance by the middle of 60th efficient power amplifiers based on both vacuum tubes and solid-state devices and great expectations for a rapid progress in their development as well as the introduction a pulse compression technique had given up the magnetrons for lost to use in the high performance radars. However, since recent time, magnetrons are considered again as a rather attractive choice to develop systems for millimeter wavelengths band namely. This turn has become possible due to: (i) a lack or low availability of other power devices operating within the indicated frequency range; (ii) a significant improvement in magnetrons characteristics, partially, incredible increase in their lifetime; (iii) a dramatic progress in digital signal processing technique; (iv) achievements in the development of high voltage modulators and millimeter wavelengths technique; and (v) a strong demand for millimeter wavelengths radars from non-military applications, which means a great interest in cost effective solutions. Despite a rather simple internal structure, the magnetron is characterized by a great complexity of the processes taken place inside it. There is no a more or less comprehensive theory of magnetrons until now. Numerical simulation can be accounted as very superficial. In general we prefer to treat the magnetron as something like a magic box characterized often by unpredictable and even surprising behavior. Thus, generally development of modern magnetron based radar, especially operating within millimeter wavelength region, requires profound understanding of principles of the magnetron operation, a great experience, and the utilization of specific design approaches, at a system level partially. However, often the magnetron is considered as old, well known device and respectively radars based on it are designed in a humble way. It results in a humble performance certainly. Probably, a rather uptight attitude to utilize the magnetrons to build contemporary high performance radars is caused by the above reason. In this paper we do not try to review comprehensively the current state of affairs or to cover as much as possible wide range of issues concerning the development of both millimeter wavelengths magnetrons and radars based on them. Instead, relying on own experience we draw attention to the fact that some noted disadvantages preventing the magnetron utilization in the high performance radar systems are essentially weakened until now, whereas others can be managed successfully by gaining from the achievements of modern electronics. The corresponding design approaches, which have assisted us to develop successfully a number of contemporary magnetron based radar systems operating within millimeter wavelength band and addressed to different application areas, will be disclosed more or less systematically. We hope that the bellow consideration will be helpful for radar designers to keep always in mind the possibilities providing by good old magnetron!

2. Magnetrons in radars – brief overview

We would not like to discuss here the physical principles of magnetron operation. They can be found in a variety of manuscripts (Okress, 1961;Tsimring, 2007). For us it is important

that the magnetron is a vacuum crossed field tube, which is capable to produce high power microwave oscillations with a high efficiency, and hence can be adopted conceptually to use in radar transmitter. As fairly noticed in (Skolnik, 2008), a choice of electronic device for the transmitter end stage defines practically completely radar structure and design approaches. Thus, let us outline the most important peculiarities of the magnetrons as related to their utilization in the radars. At first, the magnetrons are characterized fundamentally by a high both peak and average power. Typical values exceed 100 kW and 100 W for Ka band and 4 kW and 4 W for W and G bands correspondingly. It allows utilizing constant frequency pulse, being the simplest possible among radar signals, while keeping an appropriate radar sensitivity without the usage of a sophisticated signal processing technique. Next, any magnetron is an oscillator rather than an amplifier. It means partially that its output signal depends only on the physical layout of the magnetron internals and the parameters of its circumstance, i.e. applied voltage, a strength of magnetic fields, a value of voltage standing wave ratio at the output flange etc. In addition it is practically impossible to manipulate either parameter of the magnetron output signal independently from the other. Further, the magnetron is characterized by a highly resonant design basically. The magnetron oscillations frequency is essentially defined by electromagnetic properties of its internal layout and can be varied within a wide enough range only by changing a mechanical configuration of such layout, i.e. slowly. All above constrict evidently modulation capabilities providing by the magnetron. Three types of modulation are used in modern radar systems commonly – pulse (as a particular type of amplitude); frequency; and phase modulation respectively. Practically it may be considered that the magnetron by itself provides no ability for a fast, highly reproducible, and well-controlled phase/frequency modulation and adopts only the simplest pulse modulation. Probably only a bandwidth of electrical frequency chirp provided by W and G band magnetrons may appeal to use in the high resolution radars (see Section 0). Notice, since the magnetron output pulse is shaped at a radio frequency directly, it occupies it twice wider frequency band than it is required to ensure a definite spatial resolution. As for any other oscillator the magnetron oscillation frequency is subjected to fluctuations. According to common approach, fast and slow fluctuations are considered separately and referred as phase noise and frequency stability respectively. Concerning to radar performance the first defines quality of Doppler processing whereas the second is not so important generally except a number of rather special cases. Certainly, the total frequency variation range should not be too big as well as matching between transmitter and receiver frequencies should be ensured ever. Usually the related magnetron performance is characterized as poor and this fact is a byword to utilize such devices in the radars. On other hand the maximum possible variation of magnetron operational frequency including manufacturing tolerances is less than ± 1% over all millimeter wavelength bands in use even for non-tunable devices and cannot be considered as a serious issue. As for the phase noise, which is referred as a pulse-to-pulse frequency instability in the magnetron based systems, reflecting peculiarities of Doppler processor implementation, the following should be mentioned. Prima facie, it seems to be difficult actually expecting a high pulse-to-pulse frequency stability for the magnetrons. Q-factor of the magnetron resonant system is relatively low even for coaxial devices especially within millimeter wavelengths band as well as frequency pushing is rather big due to a high electron density inside an interaction space. Generally, the behavior of electron cloud has a considerable noise component and not

www.intechopen.com


well predictable at all. Initial conditions to produce successive pulses may not coincide resulting in increase in a value of the pulse-to-pulse frequency instability. This effect observed experimentally when even a single pulse altered the state of magnetron cathode surface (private communication with magnetron developers). Probably it is due to a harsh operational conditions featured by a very high peak power dissipated on as well as a high voltage applied to the elements the magnetron comprises of. On other hand, in general it is not easily to predict even very roughly what ultimate value of pulse-to-pulse frequency stability of the magnetron can be expected for each definite case. Numerical simulation is hardly useful as well as an independent direct measurement of magnetron frequency stability meets also significant difficulties if achievement of a high accuracy is mandatory. It is due to conceptually pulsed operational mode of the magnetron at rather short pulse duration as well as difficulties relating to ensure an extremely precision and reproducible shape of the high voltage pulse across the magnetron. For this reason, as we have found, despite the above seemingly evident factors, the magnetrons demonstrate a rather good performance in Doppler radar systems, certainly if appropriate design approaches are utilized (see Section 0). All above can be summarized as follows: (i) magnetron based radars operate always in a pulsed mode; (ii) a spatial resolution is determined by the duration of output magnetron pulse resulting in high peak power requirement to keep a radar potential at an appropriate level under a very low duty cycle operational condition; (iii) since each RF pulse is characterized by an arbitrary phase, a special procedure should be introduced in order to provide Doppler processing capability for the radar; and (iv) especial attention should be drawn to provide as much as possible stable magnetron operational conditions in order to achieve relevant Doppler processing performance. As related to the latter two points, it should be noticed that there is a possibility to lock the phase and frequency of magnetron oscillations with highly stable external oscillator. Unfortunately, an extremely low gain provided, i.e. relation between the output peak magnetron power and the required power of locking signal, especially for millimeter wavelengths frequency region prevent the above possibility to be used in practice. Next, magnetron life time is considered traditionally to be a serious factor limiting its usability in high performance radar systems. It reflects essentially a state of affair existing in the past, when the magnetron demonstrated actually rather low reliability caused essentially by a limited cathode lifetime. It should be noticed that the magnetron cathode operates at very high current densities as compared with other microwave tubes. In addition, the magnetron cathode is exposed strongly to electron back bombardment inherent to cross-field devices (Okress, 1961) as far as it is sited inside an interaction space. On one hand, such effect results in increase in the emission capabilities of cathode greatly due to secondary emission induced. On other hand it causes cathode overheating and affects the condition of cathode surface. It is considered that the cathode dissipates about 10 % of anode power. It means that a peak power may be as high as several kilowatts. It leads to necessity to reduce the magnetron filament power depending on a value of the anode current. The problem is such induced overheating is not well controlled and depends on many parameters. Ab initio oxide cathodes are used in the magnetrons. Due to above reasons they demonstrated a poor performance especially for millimeter wavelengths devices owing to a fine internal layout inherent to them. Our experience exposed that lifetime of W band magnetron equipped with oxide cathode was as short as several hours only! Next step was

the usage so called impregnated cathodes (Okress, 1961) in the magnetrons. It has allowed increasing in the magnetrons lifetime up to several hundred hours for Ka band devices. However despite magnetrons equipped with such cathode kept its output power and the ability to start oscillating stably within the above period, we experienced a significant frequency drift even for coaxial magnetrons caused by evaporation of the substance, from which the cathode was made of, with further absorption on the surface of magnetron cavity. Relative failures to manufacture highly reliable magnetrons have coincided with the aforementioned global drop in the interest in the development of magnetron based radars caused by other reasons. On other hand the millimeter wavelengths magnetrons are considered usually as devices for military application exclusively, and for some such applications the achieved life time seems to be more or less suitable. All above have resulted in the development of magnetron and partially investigations in cathode manufacturing have been curtailed worldwide excepting probably the former USSR. The investigations carried over there have given a new lease of life into the development of millimeter wavelengths magnetron and allowed considerable improvement of their characteristics by the end of 80th. It has been achieved due to: (i) utilization of metallic alloy cathodes; (ii) successes in the development of the magnetrons with cold secondary emission cathodes; and (iii) utilization of spatial harmonics different from type. The latter was especially important for W and G band magnetrons since it has allowed to enlarge dimensions of the magnetron interaction space and the cathode diameter, which results in a considerable increase in peak and average power as well as the maximal pulse duration. In addition, the usage of samarium cobalt magnet system has allowed reducing the magnetron dimensions and weight as well as developing rather miniature devices. The parameters of several millimeter wavelengths magnetrons developed with utilization of above approaches are summarized in Table I.

Frequency band Ka Ka W G Peak output power, kW

50 3 4 10

Maximal duty cycle, %

0.2 0.1 0.1 0.085

Anode voltage, V 14000 6300 10000 5 Efficiency > 33% > 20% > 4% > 5% Type Coaxial No information Spatial Harmonic Spatial Harmonic Type of cathode Metallic alloy Cold with spike

autoemmiters Cold with auxiliary thermionic cathode

Metallic alloy (?)

Frequency agility Yes No No No Life time, hours Producer specified Reached during utilization Average Maximal

>2000

>10000 >30000

>2000

>1000

>5000 >10000

No information

Cooling Forced-air Forced-air Forced-air Liquid Production status Full

production Full production Pre production Full production

Table I. Parameters of millimeter wavelengths magnetrons.

www.intechopen.com


well predictable at all. Initial conditions to produce successive pulses may not coincide resulting in increase in a value of the pulse-to-pulse frequency instability. This effect observed experimentally when even a single pulse altered the state of magnetron cathode surface (private communication with magnetron developers). Probably it is due to a harsh operational conditions featured by a very high peak power dissipated on as well as a high voltage applied to the elements the magnetron comprises of. On other hand, in general it is not easily to predict even very roughly what ultimate value of pulse-to-pulse frequency stability of the magnetron can be expected for each definite case. Numerical simulation is hardly useful as well as an independent direct measurement of magnetron frequency stability meets also significant difficulties if achievement of a high accuracy is mandatory. It is due to conceptually pulsed operational mode of the magnetron at rather short pulse duration as well as difficulties relating to ensure an extremely precision and reproducible shape of the high voltage pulse across the magnetron. For this reason, as we have found, despite the above seemingly evident factors, the magnetrons demonstrate a rather good performance in Doppler radar systems, certainly if appropriate design approaches are utilized (see Section 0). All above can be summarized as follows: (i) magnetron based radars operate always in a pulsed mode; (ii) a spatial resolution is determined by the duration of output magnetron pulse resulting in high peak power requirement to keep a radar potential at an appropriate level under a very low duty cycle operational condition; (iii) since each RF pulse is characterized by an arbitrary phase, a special procedure should be introduced in order to provide Doppler processing capability for the radar; and (iv) especial attention should be drawn to provide as much as possible stable magnetron operational conditions in order to achieve relevant Doppler processing performance. As related to the latter two points, it should be noticed that there is a possibility to lock the phase and frequency of magnetron oscillations with highly stable external oscillator. Unfortunately, an extremely low gain provided, i.e. relation between the output peak magnetron power and the required power of locking signal, especially for millimeter wavelengths frequency region prevent the above possibility to be used in practice. Next, magnetron life time is considered traditionally to be a serious factor limiting its usability in high performance radar systems. It reflects essentially a state of affair existing in the past, when the magnetron demonstrated actually rather low reliability caused essentially by a limited cathode lifetime. It should be noticed that the magnetron cathode operates at very high current densities as compared with other microwave tubes. In addition, the magnetron cathode is exposed strongly to electron back bombardment inherent to cross-field devices (Okress, 1961) as far as it is sited inside an interaction space. On one hand, such effect results in increase in the emission capabilities of cathode greatly due to secondary emission induced. On other hand it causes cathode overheating and affects the condition of cathode surface. It is considered that the cathode dissipates about 10 % of anode power. It means that a peak power may be as high as several kilowatts. It leads to necessity to reduce the magnetron filament power depending on a value of the anode current. The problem is such induced overheating is not well controlled and depends on many parameters. Ab initio oxide cathodes are used in the magnetrons. Due to above reasons they demonstrated a poor performance especially for millimeter wavelengths devices owing to a fine internal layout inherent to them. Our experience exposed that lifetime of W band magnetron equipped with oxide cathode was as short as several hours only! Next step was

the usage so called impregnated cathodes (Okress, 1961) in the magnetrons. It has allowed increasing in the magnetrons lifetime up to several hundred hours for Ka band devices. However despite magnetrons equipped with such cathode kept its output power and the ability to start oscillating stably within the above period, we experienced a significant frequency drift even for coaxial magnetrons caused by evaporation of the substance, from which the cathode was made of, with further absorption on the surface of magnetron cavity. Relative failures to manufacture highly reliable magnetrons have coincided with the aforementioned global drop in the interest in the development of magnetron based radars caused by other reasons. On other hand the millimeter wavelengths magnetrons are considered usually as devices for military application exclusively, and for some such applications the achieved life time seems to be more or less suitable. All above have resulted in the development of magnetron and partially investigations in cathode manufacturing have been curtailed worldwide excepting probably the former USSR. The investigations carried over there have given a new lease of life into the development of millimeter wavelengths magnetron and allowed considerable improvement of their characteristics by the end of 80th. It has been achieved due to: (i) utilization of metallic alloy cathodes; (ii) successes in the development of the magnetrons with cold secondary emission cathodes; and (iii) utilization of spatial harmonics different from type. The latter was especially important for W and G band magnetrons since it has allowed to enlarge dimensions of the magnetron interaction space and the cathode diameter, which results in a considerable increase in peak and average power as well as the maximal pulse duration. In addition, the usage of samarium cobalt magnet system has allowed reducing the magnetron dimensions and weight as well as developing rather miniature devices. The parameters of several millimeter wavelengths magnetrons developed with utilization of above approaches are summarized in Table I.

Frequency band Ka Ka W G Peak output power, kW

50 3 4 10

Maximal duty cycle, %

0.2 0.1 0.1 0.085

Anode voltage, V 14000 6300 10000 5 Efficiency > 33% > 20% > 4% > 5% Type Coaxial No information Spatial Harmonic Spatial Harmonic Type of cathode Metallic alloy Cold with spike

autoemmiters Cold with auxiliary thermionic cathode

Metallic alloy (?)

Frequency agility Yes No No No Life time, hours Producer specified Reached during utilization Average Maximal

>2000

>10000 >30000

>2000

>1000

>5000 >10000

No information

Cooling Forced-air Forced-air Forced-air Liquid Production status Full

production Full production Pre production Full production

Table I. Parameters of millimeter wavelengths magnetrons.

www.intechopen.com


The indicated values for lifetime of Ka band magnetrons have been obtained during the utilization of a line of Ka band meteorological radars (see Section 0), whereas W-band magnetrons were tested in course of a particular procedure. Achieved value of the magnetron lifetime can be qualified literally as outstanding! It should be noticed that above results were absolutely unexpected for us – we considered reaching 4000 hours only in the best case for Ka band magnetrons. There were no selection for Ka band magnetrons – either among them demonstrates similar performance. Why it has become possible? Beside of the above advances in the cathode manufacturing the overall improvement in production technology should Be mentioned, which ensure keeping a high vacuum condition inside the magnetron during whole utilization term. In addition we dare to claim that an advanced design of a magnetron modulator helps significantly the magnetrons exposing its actual potential by automating providing as much as possible safe and optimal operational mode. Since W band magnetrons in question are manufactured in one of branch of Institute of Radio Astronomy of National Academy of Sciences of Ukraine we would like to describe their development and production in some more detail. The first millimeter wavelengths devices had been developed in the middle of 60th (Usikov, 1972). Ab initio they utilize operation at spatial harmonics different from type. Such magnetron demonstrates excellent performance e.g. peak output power achieved was 80 kW and 10 kW for 3 mm and 2 mm wavelengths devices correspondingly. However due to L type oxide cathode was used in the magnetrons, their life time was limited by a value of several tens hours only. In order to overcome this problem, a cold secondary emission cathode has been introduced in the magnetron design. An auxiliary thermionic cathode placed aside an interaction space was used to provide an initial electron density in the magnetron and ensure oscillation running at the front of modulation pulse. Despite such magnetrons demonstrates inherently a lower efficiency, they are much more promising to extend their lifetime. As a result W band 1 kW device has been developed and industrialized by the middle of 80th (Naumenko et al, 1999). It characterized by a guaranteed life time of 2000 hours. By the end of 90th essential efforts has been concentrated to develop W band magnetrons with expected life time of several thousand hours at peak power of 4 kW, pulse duration of 200 nsec, and duty cycle of 0.1 % for meteorological radars and, a bit later, 1 kW devices with target life time of 10000 hours for airport debris radar. These efforts have resulted in a stable output of W band magnetrons characterized by peak power within the range from 1000 to 4000 W and expected lifetime of 10000 hours (Gritsaenko et al, 2005). At the time being every third manufactured magnetron satisfies technical specification and there are serious reasons to expect an improvement of this ratio. Such fact can be considered as a serious claim on further industrialization. The following problems have been solved to extend the magnetron life time: (i) optimal design of the auxiliary thermionic cathode; (ii) correct choice of an operational spatial harmonics; and (iii) equipping the magnetron with a magnetic-discharge vacuum pump. The latter increases the magnetron dimension but is mandatory to ensure its long life operation especially at a heavy operational conditions, such as a long pulse width, a large duty cycle, or a high pulse repetition rate (see Section 0). A photo of 4 kW W band magnetron produced in the Institute of Radio Astronomy is depicted in Fig. 1. During resent time a low voltage, compact Ka band magnetrons with cold secondary emission cathode are under

Fig. 1. 4 kW, W band magnetron with cold secondary emission cathode and extended life time. development. They are intended to be used in low-cost meteorological radar sensors, which are especially convenient for network applications. Above we have considered the peculiarities of the magnetron utilization in the radars as well as demonstrated that the life time is not an issue preventing magnetrons from usage in high performance radars. Now we would like to discuss further benefits and disadvantages of such approach in a comparative manner. Table II provides a brief comparison between some high power millimeter wavelengths devices available at the time being in respect to their possible usage in the radars. A comprehensive review of current state of the development of power millimeter wavelengths sources can be found in (Barker et al, 2005). Some remarks should be done in addition. As usual the possibility to introduce a sophisticated signal modulation is mentioned as the essential benefit provided by using an amplifier in the radar transmitter. Actually the utilization of a pulse compression technique allows attaining the highest possible resolution for the radars, which is of centimeters grade now. Certainly, the magnetron based radars cannot provide a similar performance level. However for the most applications an extreme resolution is not required and pulse compression is used only to ensure suitable radar sensitivity.

Magnetron Traveling wave tube

Klystron with extended interaction

Solid state power amplifier

Peak (average) output power, kW(W) Ka band W band

70 (100)1 4 (4)1

1 (300) 1 [50 (500)]2 1 (10) 1 [3 (300)]2

3 (300) 1 1.5 (90) 1 [1.5 (150)]

2

0.004 (0.004) (single chip) 0.0002 (0.0002) (single chip)

Type Oscillator Amplifier Amplifier Amplifier Resolution achievable High Highest Highest Highest Life time Long Long Long Long Cost Low High High Low (single chip) Mandatory options of signal processing

Coherence on receiver

Pulse compression

Pulse compression Pulse compression; Power combining

Total system cost at similar specifications

Low/moderate High High Very High

Table II. Parameters of power microwave devices.

1 Air or conductive cooling 2 Water cooling

www.intechopen.com


The indicated values for lifetime of Ka band magnetrons have been obtained during the utilization of a line of Ka band meteorological radars (see Section 0), whereas W-band magnetrons were tested in course of a particular procedure. Achieved value of the magnetron lifetime can be qualified literally as outstanding! It should be noticed that above results were absolutely unexpected for us – we considered reaching 4000 hours only in the best case for Ka band magnetrons. There were no selection for Ka band magnetrons – either among them demonstrates similar performance. Why it has become possible? Beside of the above advances in the cathode manufacturing the overall improvement in production technology should Be mentioned, which ensure keeping a high vacuum condition inside the magnetron during whole utilization term. In addition we dare to claim that an advanced design of a magnetron modulator helps significantly the magnetrons exposing its actual potential by automating providing as much as possible safe and optimal operational mode. Since W band magnetrons in question are manufactured in one of branch of Institute of Radio Astronomy of National Academy of Sciences of Ukraine we would like to describe their development and production in some more detail. The first millimeter wavelengths devices had been developed in the middle of 60th (Usikov, 1972). Ab initio they utilize operation at spatial harmonics different from type. Such magnetron demonstrates excellent performance e.g. peak output power achieved was 80 kW and 10 kW for 3 mm and 2 mm wavelengths devices correspondingly. However due to L type oxide cathode was used in the magnetrons, their life time was limited by a value of several tens hours only. In order to overcome this problem, a cold secondary emission cathode has been introduced in the magnetron design. An auxiliary thermionic cathode placed aside an interaction space was used to provide an initial electron density in the magnetron and ensure oscillation running at the front of modulation pulse. Despite such magnetrons demonstrates inherently a lower efficiency, they are much more promising to extend their lifetime. As a result W band 1 kW device has been developed and industrialized by the middle of 80th (Naumenko et al, 1999). It characterized by a guaranteed life time of 2000 hours. By the end of 90th essential efforts has been concentrated to develop W band magnetrons with expected life time of several thousand hours at peak power of 4 kW, pulse duration of 200 nsec, and duty cycle of 0.1 % for meteorological radars and, a bit later, 1 kW devices with target life time of 10000 hours for airport debris radar. These efforts have resulted in a stable output of W band magnetrons characterized by peak power within the range from 1000 to 4000 W and expected lifetime of 10000 hours (Gritsaenko et al, 2005). At the time being every third manufactured magnetron satisfies technical specification and there are serious reasons to expect an improvement of this ratio. Such fact can be considered as a serious claim on further industrialization. The following problems have been solved to extend the magnetron life time: (i) optimal design of the auxiliary thermionic cathode; (ii) correct choice of an operational spatial harmonics; and (iii) equipping the magnetron with a magnetic-discharge vacuum pump. The latter increases the magnetron dimension but is mandatory to ensure its long life operation especially at a heavy operational conditions, such as a long pulse width, a large duty cycle, or a high pulse repetition rate (see Section 0). A photo of 4 kW W band magnetron produced in the Institute of Radio Astronomy is depicted in Fig. 1. During resent time a low voltage, compact Ka band magnetrons with cold secondary emission cathode are under

Fig. 1. 4 kW, W band magnetron with cold secondary emission cathode and extended life time. development. They are intended to be used in low-cost meteorological radar sensors, which are especially convenient for network applications. Above we have considered the peculiarities of the magnetron utilization in the radars as well as demonstrated that the life time is not an issue preventing magnetrons from usage in high performance radars. Now we would like to discuss further benefits and disadvantages of such approach in a comparative manner. Table II provides a brief comparison between some high power millimeter wavelengths devices available at the time being in respect to their possible usage in the radars. A comprehensive review of current state of the development of power millimeter wavelengths sources can be found in (Barker et al, 2005). Some remarks should be done in addition. As usual the possibility to introduce a sophisticated signal modulation is mentioned as the essential benefit provided by using an amplifier in the radar transmitter. Actually the utilization of a pulse compression technique allows attaining the highest possible resolution for the radars, which is of centimeters grade now. Certainly, the magnetron based radars cannot provide a similar performance level. However for the most applications an extreme resolution is not required and pulse compression is used only to ensure suitable radar sensitivity.

Magnetron Traveling wave tube

Klystron with extended interaction

Solid state power amplifier

Peak (average) output power, kW(W) Ka band W band

70 (100)1 4 (4)1

1 (300) 1 [50 (500)]2 1 (10) 1 [3 (300)]2

3 (300) 1 1.5 (90) 1 [1.5 (150)]

2

0.004 (0.004) (single chip) 0.0002 (0.0002) (single chip)

Type Oscillator Amplifier Amplifier Amplifier Resolution achievable High Highest Highest Highest Life time Long Long Long Long Cost Low High High Low (single chip) Mandatory options of signal processing

Coherence on receiver

Pulse compression

Pulse compression Pulse compression; Power combining

Total system cost at similar specifications

Low/moderate High High Very High

Table II. Parameters of power microwave devices.

1 Air or conductive cooling 2 Water cooling

www.intechopen.com


In this case an inherent adverse effect of pulse compression utilization, namely, appearance of spurious targets or, instead, masquerading low RCS targets in presence of that with a strong reflectivity due to side-lobes of the autocorrelation function, should be always taken into consideration. For example, the necessity to monitor simultaneously meteorological phenomena (clouds, precipitation) with a reflectivity range of about 100 dB restricts essentially the application of pulse compression technique for weather radars. Certainly the utilization of such technique providing a relevant performance level requires much more complicated signal processing hardware. Relatively inexpensive implementation of such hardware, suitable to use in mainstream commercial radars, has become to be allowable since very recent time. Next issue, mentioned usually to accent benefits of truly coherent systems based on utilizing an amplifier in the transmitter, is much higher quality of Doppler processing in this case. However recently developed magnetron based radar systems (see Section 3) demonstrates a competitive Doppler performance within at least Ka band where coaxial magnetrons are available. In addition, further advances in the magnetron modulator design as well as the utilization of a sophisticated digital signal processing allows us to expect over and above improvements in this area, especially for W band radars. The accuracy achieved currently for the phase processing in the magnetron based radars allows us to suggest retrieving appropriate performance even for applications requiring the utilization of synthetic aperture. Interestingly, the approach, based on sampling the radiated signal to provide an enhanced signal processing, inherent to the magnetron based radars, can be useful to improve processing of compressed pulses by avoiding the influence of the distortions introduced by Tx/Rx chains (Zhu, 2008). It is an extra sign that signal processing approaches and, then, capabilities providing by both truly and pseudo coherent radars become closer. Probably the only area the magnetron based radars cannot compete beyond any doubts with truly coherent systems is radars requiring fast frequency agility. It should be noticed that despite a low power produced by a single chip, solid state devices allows conceptually developing the most advanced radar systems ever with an unprecedented performance level due to utilization of active phased array technique. Certainly, within millimeter wavelength band such approach is on the technology edge currently and requires enormous efforts to be implemented. All above mentioned allow us to declare that the achievable performance level of the magnetron based radars is appropriate to meet the most requirements for mainstream applications. Their relatively low cost and complexity make them being very attractive solution especially in the case if the radar system benefits essentially from a shorter wavelengths, as for meteorological applications (1/2 law for the reflectivity of meteors). Applications where dimensions and weight is among major requirements can be considered also as the area of preferable utilization of the magnetron based radars especially operating within W and G bands.

3. Magnetron based radar systems – development experience

Development of the magnetron based radar systems had started in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine since the middle of 90th of the last century. By the time indicated we experienced with success during many years in the

development and manufacturing of millimetre wavelengths magnetrons with cold secondary emission cathode (Gritsaenko et al, 2005). At the same time, it was clear that existing circumstance for the magnetron usage did not meet modern requirements and did not allow disclosing an actual potential of such devices in radar applications. Then considerable efforts were made in order to develop advanced modulators to drive the above magnetrons, which require furthermore a tighter control of modulation pulse shape as compared to traditional types. Such situation has coincided with a growing interest in radars for environmental investigations operating within millimeter wavelength band. In the course of above tendency the first magnetron based radar sensor has been developed and tested by us (Jenett et all, 1999). It was a rather simple Ka and W band double frequency airborne side looking system intended to detect oil spills on water surface. Since this task benefits strongly from the wavelengths shortening, the utilization of magnetrons has allowed developing the radar rapidly by using simple non coherent signal processing. Nevertheless it demonstrated unexpectedly good performance and was capable to detect a rather weak oil spills even under a low waves condition inherent for internal reservoirs. Obtained experience has allowed us to proceed with radar development and, naturally, the next step was the introduction of Doppler processing capabilities in the radar (Schunemann et al, 2000). After the first usage of a traditional analogue coherence-on-receiver technique, its digital implementation has been introduced. Experiments with a prototype of Ka band meteorological radar demonstrated a good Doppler performance, which was appropriate for the most atmospheric researches as well as to monitor atmospheric conditions. In addition, a capability of the magnetron to work continuously during over several thousand hours has been proven. However our first full functional Doppler polarimetric meteorological radar was equipped with two magnetron based transmitters in order to ensure its reliable unattended continuous operation for at least several months interval. The experience of first year utilization of this radar has disclosed a surprisingly high stability of the magnetron operation. It has allowed to lead off developing a line of high performance magnetron based meteorological radars. It includes both vertically pointed and scanning systems with a high mobility as depicted in Fig. 2. Until now

Fig. 2. Ka band magnetron based meteorological radars. seven radars has been produced and delivered in cooperation with METEK GmbH (Elmshorn, Germany). Some of them are included into European weather radar network.

www.intechopen.com


In this case an inherent adverse effect of pulse compression utilization, namely, appearance of spurious targets or, instead, masquerading low RCS targets in presence of that with a strong reflectivity due to side-lobes of the autocorrelation function, should be always taken into consideration. For example, the necessity to monitor simultaneously meteorological phenomena (clouds, precipitation) with a reflectivity range of about 100 dB restricts essentially the application of pulse compression technique for weather radars. Certainly the utilization of such technique providing a relevant performance level requires much more complicated signal processing hardware. Relatively inexpensive implementation of such hardware, suitable to use in mainstream commercial radars, has become to be allowable since very recent time. Next issue, mentioned usually to accent benefits of truly coherent systems based on utilizing an amplifier in the transmitter, is much higher quality of Doppler processing in this case. However recently developed magnetron based radar systems (see Section 3) demonstrates a competitive Doppler performance within at least Ka band where coaxial magnetrons are available. In addition, further advances in the magnetron modulator design as well as the utilization of a sophisticated digital signal processing allows us to expect over and above improvements in this area, especially for W band radars. The accuracy achieved currently for the phase processing in the magnetron based radars allows us to suggest retrieving appropriate performance even for applications requiring the utilization of synthetic aperture. Interestingly, the approach, based on sampling the radiated signal to provide an enhanced signal processing, inherent to the magnetron based radars, can be useful to improve processing of compressed pulses by avoiding the influence of the distortions introduced by Tx/Rx chains (Zhu, 2008). It is an extra sign that signal processing approaches and, then, capabilities providing by both truly and pseudo coherent radars become closer. Probably the only area the magnetron based radars cannot compete beyond any doubts with truly coherent systems is radars requiring fast frequency agility. It should be noticed that despite a low power produced by a single chip, solid state devices allows conceptually developing the most advanced radar systems ever with an unprecedented performance level due to utilization of active phased array technique. Certainly, within millimeter wavelength band such approach is on the technology edge currently and requires enormous efforts to be implemented. All above mentioned allow us to declare that the achievable performance level of the magnetron based radars is appropriate to meet the most requirements for mainstream applications. Their relatively low cost and complexity make them being very attractive solution especially in the case if the radar system benefits essentially from a shorter wavelengths, as for meteorological applications (1/2 law for the reflectivity of meteors). Applications where dimensions and weight is among major requirements can be considered also as the area of preferable utilization of the magnetron based radars especially operating within W and G bands.

3. Magnetron based radar systems – development experience

Development of the magnetron based radar systems had started in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine since the middle of 90th of the last century. By the time indicated we experienced with success during many years in the

development and manufacturing of millimetre wavelengths magnetrons with cold secondary emission cathode (Gritsaenko et al, 2005). At the same time, it was clear that existing circumstance for the magnetron usage did not meet modern requirements and did not allow disclosing an actual potential of such devices in radar applications. Then considerable efforts were made in order to develop advanced modulators to drive the above magnetrons, which require furthermore a tighter control of modulation pulse shape as compared to traditional types. Such situation has coincided with a growing interest in radars for environmental investigations operating within millimeter wavelength band. In the course of above tendency the first magnetron based radar sensor has been developed and tested by us (Jenett et all, 1999). It was a rather simple Ka and W band double frequency airborne side looking system intended to detect oil spills on water surface. Since this task benefits strongly from the wavelengths shortening, the utilization of magnetrons has allowed developing the radar rapidly by using simple non coherent signal processing. Nevertheless it demonstrated unexpectedly good performance and was capable to detect a rather weak oil spills even under a low waves condition inherent for internal reservoirs. Obtained experience has allowed us to proceed with radar development and, naturally, the next step was the introduction of Doppler processing capabilities in the radar (Schunemann et al, 2000). After the first usage of a traditional analogue coherence-on-receiver technique, its digital implementation has been introduced. Experiments with a prototype of Ka band meteorological radar demonstrated a good Doppler performance, which was appropriate for the most atmospheric researches as well as to monitor atmospheric conditions. In addition, a capability of the magnetron to work continuously during over several thousand hours has been proven. However our first full functional Doppler polarimetric meteorological radar was equipped with two magnetron based transmitters in order to ensure its reliable unattended continuous operation for at least several months interval. The experience of first year utilization of this radar has disclosed a surprisingly high stability of the magnetron operation. It has allowed to lead off developing a line of high performance magnetron based meteorological radars. It includes both vertically pointed and scanning systems with a high mobility as depicted in Fig. 2. Until now

Fig. 2. Ka band magnetron based meteorological radars. seven radars has been produced and delivered in cooperation with METEK GmbH (Elmshorn, Germany). Some of them are included into European weather radar network.

www.intechopen.com


Coaxial magnetrons are used in the radars (see Table I). Many improvements have been brought into design with every new item including a double frequency conversion in the receiver; a digital automatic frequency control; a digital receiver technique implementation; modifications in receiver protection circuitry; introduction of the circuits ensuring more magnetron operation safety etc. The essential parameters of most resent radars are summarized in Table III. The quality of Doppler processing provided by such radars is illustrated by Fig. 3 where Doppler spectrum obtained from a stationary target located at the distance of about 5.5 km is depicted. Signal processing parameters were as follows: pulse repetition frequency – 10 kHz; fast Fourier transform length – 512; spectrum averaging – 10 (dwell time - 0.5 sec). As follows from this figure, Doppler dynamical range exceeds 60 dBc, which corresponding to the value of wideband noise floor of -73 dBc/Hz. From this data it is possible to estimate a value of the magnetron pulse-to-pulse frequency instability. Actually, the total power of the received signal can be considered as a sum of coherent �� and incoherent �� components respectively. According to general principles of Doppler processing in radars (Skolnik, 2008), taking into account that the above data are product of a discrete Fourier transform (DFT), and assuming that the magnetron introduces a noise distributed evenly in frequency domain, the ratio between the above components for the signal backscattered by a stationary target, producing definitely monochromatic response, can be set down as:

�� · ��, (1)

where �� is the noise floor of the radar Doppler processing and �� is pulse repetition frequency of the radar. On other hand, the phase lag of the signal reflected from a stationary target located at a fixed distance R is �� , where �� is the deviation of the wavelengths for ith pulse from a constant value of ��. Assuming that �� , the corresponding discrete time complex signal S at the input of DFT may be written as follows:

�� · �� · �� · �� · �� · �� , (2)

where �� . The second term in the above equation reflects the entity of incoherent components in the received signal due to the magnetron pulse-to-pulse frequency instability and determines the value of noise floor �� on

Operating frequency, GHz 35.5±0.15 Peak transmitter power, kW 30 Average power (max), W 50 Losses, dB, Tx path Rx path

1 2.5

Pulse duration, nsec 100, 200, 400 Pulse repetition rate, kHz 5…10 Receiver noise figure, dB 3.0 Radar instantaneous dynamical range including STC, dB >80 Distance for sensitivity, m at -10 dB at -1 dB

180 330

Ta Donema

SuPRof

Figban An(Vaplinlanwh

AntenPolarTotalDoppNumVolumTransReceiWeig

able III. Parameter

oppler spectrum agligible. Thus fagnetron pulse-to

ubstitution of varRF of 10 kHz, o

about .

g. 3. Doppler specnd meteorologica

nother type of a colkov et al, 20plications related

nes and other obnding. But the achich can be used

nna gain rization decoupling impressed compo

pler processing @ 5 mber of range bins

me smitter iver

ght of Tx/Rx units, krs of Ka band me

assuming that thefinally from formo-pulse frequency

riables in (3) withof 8.2 mm, and R

ctrum from statioal radar.

compact magnetr007).The developd to enhancing hbstacles, monitorhieved radar per

d for other appli

g osite wideband no km, dBc/Hz

kg teorological rada

e contribution fromulas (1) and y instability reads

h the above valuR of 5.5 km result

onary target locat

ron based Ka banped radar systemhelicopter flight sring meteorologrformance enablecations. The rad

>5>4

oise of -5

4

ars.

om other sources (2), the followins as:

ues, namely, noiss in pulse-to-puls

ted at 5.5 km dist

nd radar is airborm has been desafety including tical conditions,

es us to consider ar benefits from

50 dB 40 dB

-70 500

9U 4U 40

like a local oscillng expression fo

se floor of -73 dBse frequency inst

tance retrieved w

rne multipurposeesigned especialthe detection of and providing it as a versatile s

m some novel and

lator is or the

(3)

Bc/Hz, tability

with Ka

e radar lly for power secure

sensor, d cost-

www.intechopen.com


Coaxial magnetrons are used in the radars (see Table I). Many improvements have been brought into design with every new item including a double frequency conversion in the receiver; a digital automatic frequency control; a digital receiver technique implementation; modifications in receiver protection circuitry; introduction of the circuits ensuring more magnetron operation safety etc. The essential parameters of most resent radars are summarized in Table III. The quality of Doppler processing provided by such radars is illustrated by Fig. 3 where Doppler spectrum obtained from a stationary target located at the distance of about 5.5 km is depicted. Signal processing parameters were as follows: pulse repetition frequency – 10 kHz; fast Fourier transform length – 512; spectrum averaging – 10 (dwell time - 0.5 sec). As follows from this figure, Doppler dynamical range exceeds 60 dBc, which corresponding to the value of wideband noise floor of -73 dBc/Hz. From this data it is possible to estimate a value of the magnetron pulse-to-pulse frequency instability. Actually, the total power of the received signal can be considered as a sum of coherent �� and incoherent �� components respectively. According to general principles of Doppler processing in radars (Skolnik, 2008), taking into account that the above data are product of a discrete Fourier transform (DFT), and assuming that the magnetron introduces a noise distributed evenly in frequency domain, the ratio between the above components for the signal backscattered by a stationary target, producing definitely monochromatic response, can be set down as:

�� · ��, (1)

where �� is the noise floor of the radar Doppler processing and �� is pulse repetition frequency of the radar. On other hand, the phase lag of the signal reflected from a stationary target located at a fixed distance R is �� , where �� is the deviation of the wavelengths for ith pulse from a constant value of ��. Assuming that �� , the corresponding discrete time complex signal S at the input of DFT may be written as follows:

�� · �� · �� · �� · �� · �� , (2)

where �� . The second term in the above equation reflects the entity of incoherent components in the received signal due to the magnetron pulse-to-pulse frequency instability and determines the value of noise floor �� on

Operating frequency, GHz 35.5±0.15 Peak transmitter power, kW 30 Average power (max), W 50 Losses, dB, Tx path Rx path

1 2.5

Pulse duration, nsec 100, 200, 400 Pulse repetition rate, kHz 5…10 Receiver noise figure, dB 3.0 Radar instantaneous dynamical range including STC, dB >80 Distance for sensitivity, m at -10 dB at -1 dB

180 330

Ta Donema

SuPRof

Figban An(Vaplinlanwh

AntenPolarTotalDoppNumVolumTransReceiWeig

able III. Parameter

oppler spectrum agligible. Thus fagnetron pulse-to

ubstitution of varRF of 10 kHz, o

about .

g. 3. Doppler specnd meteorologica

nother type of a colkov et al, 20plications related

nes and other obnding. But the achich can be used

nna gain rization decoupling impressed compo

pler processing @ 5 mber of range bins

me smitter iver

ght of Tx/Rx units, krs of Ka band me

assuming that thefinally from formo-pulse frequency

riables in (3) withof 8.2 mm, and R

ctrum from statioal radar.

compact magnetr007).The developd to enhancing hbstacles, monitorhieved radar per

d for other appli

g osite wideband no km, dBc/Hz

kg teorological rada

e contribution fromulas (1) and y instability reads

h the above valuR of 5.5 km result

onary target locat

ron based Ka banped radar systemhelicopter flight sring meteorologrformance enablecations. The rad

>5>4

oise of -5

4

ars.

om other sources (2), the followins as:

ues, namely, noiss in pulse-to-puls

ted at 5.5 km dist

nd radar is airborm has been desafety including tical conditions,

es us to consider ar benefits from

50 dB 40 dB

-70 500

9U 4U 40

like a local oscillng expression fo

se floor of -73 dBse frequency inst

tance retrieved w

rne multipurposeesigned especialthe detection of and providing it as a versatile s

m some novel and

lator is or the

(3)

Bc/Hz, tability

with Ka

e radar lly for power secure

sensor, d cost-

www.intechopen.com


effwasysFig

Figrad corcohproweess

Ta

fective solutions aveguide antennastem. The radar og. 4 a, b, and c

a) g. 4. Outline (a), bdar.

rrespondingly. Inherence-on-receivotection circuitryell as a local oscilsential radar para

Operating Peak transmLosses, dB,

Pulse duraPulse repetReceiver n

Receiver dyTime-variaMinimal di

Antenna Antenna pAntenna seAntenna seAntenna swSingle-pulsTotal impr@ 3 km, dBPrimary poPower con

Volume, litWeight, kg

able IV. Parameter

including a low-a array as well as outline; a simplif

block diagram (b

n the radar there ver implementat

y is used for this lator based on a d

ameters are summ

frequency, GHz mitter power, kW , Tx path Rx path

ation, nsec tition rate, kHz oise figure, dB

ynamical range (max)ation gain control rangistance for full sensitiv

olarization ection switching ection switching time,witch decoupling, dB se sensitivity at 2 km, ressed composite wideBc/Hz ower supply, V

nsumption (max), W

tre g rs of airborne Ka

-noise; digital rec a multifunctionafied block diagra

b) b), and antenna p

is no a separate ction. Instead the purpose (see Secdirect digital syn

marized in Table I

), dB ge vity, m

, µsec

dBm2 eband noise of Dopple

band scanning ra

ceiver; an electrical data acquisitionam; and antenna

pattern (c) of Ka b

channel to samplee signal leaked ction 0). Single fr

nthesizer has beenIV.

er processing

adar

cally switchable sn and signal procpattern are depic

c) band airborne sca

e the radiated pu through the rerequency conversn used in the rada

35±0.2 2.5

1.5 2.5

50…500 1…10

3.0 >90 dB 24 dB

50 4-sections

vertical electrical

1 > 25 -13

-53

18-32 DC 70 12 12

slotted cessing cted in

anning

ulse for eceiver sion as ar. The

A prototype of meteorological W band radar has been developed in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine (Vavriv et al, 2002). It utilizes a proprietary magnetron with cold secondary emission cathode. (see Table I). It is featured by: (i) two separate antennas; (ii) a separate downconverter to sample the radiated signal; and (iii) a high power quasi-optical polarization rotators both in Tx and Rx channels. The radar characteristics are summarized in Table V. It should be noticed that the radar in question has been developed several years ago and unfortunately no modifications have been provided till now due to our effort were concentrated on the radars operating within Ka band essentially.

Operating frequency, GHz 94 Peak power (max), kW 4 Pulse width, ns 50 - 400 Pulse repetition frequency, kHz 2.5 - 10 Receiver noise temperature, K 1200 Total dynamical range, dB 70 Polarization HH, VV, HV, VH Cross-polarization isolation, dB -25 Antenna diameter, m 0.5 Antenna beam width, deg 0.45 Total impressed composite wideband noise of Doppler processing @ 5 km, dBc/Hz -47

Sensitivity at 5 km with the integration time of 0.1 sec, dBZ -41

Table V. Parameters of prototype of W band meteorological radar. Therefore a serious improvement of its parameter may be expected due to: (i) utilization of a single antenna due to increase in availability of high power circulators and P-i-N switches for the receiver protection circuitry; (ii) an introduction of digital receiver technique as well as a digital frequency control similarly to the above Ka band radars; (iii) the introduction of a low noise amplifier, which have become available during resent time; (iv) the introduction of a synthesized local oscillator; and (v) the introduction of an advanced magnetron modulator. And at last we would like to provide a brief description W band short pulse transmitter for airport debris radar (Belikov et al, 2002). As known debris are very serious problem to provide enough flight safety. A high resolution as well as a high reliability is a mandatory requirement conceptually for such radars. In the case of magnetron based radar an appropriate resolution may be achieved in a rather simple way, without the utilization of pulse compression. The magnetron with cold-secondary emission cathode used in the transmitter in question provides in addition an extended life time of 10000 hours at least. The parameters of the transmitter are given in Table VI.

Parameter Measure Unit Measured Value Output Frequency(25°C) GHz 95.086 Frequency Temperature Coefficient MHz/°C -1.75 RF Pulse Width (@-3dB, 25°C) ns 17.5 RF Pulse Width (@-6dB, 25°C) ns 20 RF Output Peak Power kW 2

www.intechopen.com


effwasysFig

Figrad corcohproweess

Ta

fective solutions aveguide antennastem. The radar og. 4 a, b, and c

a) g. 4. Outline (a), bdar.

rrespondingly. Inherence-on-receivotection circuitryell as a local oscilsential radar para

Operating Peak transmLosses, dB,

Pulse duraPulse repetReceiver n

Receiver dyTime-variaMinimal di

Antenna Antenna pAntenna seAntenna seAntenna swSingle-pulsTotal impr@ 3 km, dBPrimary poPower con

Volume, litWeight, kg

able IV. Parameter

including a low-a array as well as outline; a simplif

block diagram (b

n the radar there ver implementat

y is used for this lator based on a d

ameters are summ

frequency, GHz mitter power, kW , Tx path Rx path

ation, nsec tition rate, kHz oise figure, dB

ynamical range (max)ation gain control rangistance for full sensitiv

olarization ection switching ection switching time,witch decoupling, dB se sensitivity at 2 km, ressed composite wideBc/Hz ower supply, V

nsumption (max), W

tre g rs of airborne Ka

-noise; digital rec a multifunctionafied block diagra

b) b), and antenna p

is no a separate ction. Instead the purpose (see Secdirect digital syn

marized in Table I

), dB ge vity, m

, µsec

dBm2 eband noise of Dopple

band scanning ra

ceiver; an electrical data acquisitionam; and antenna

pattern (c) of Ka b

channel to samplee signal leaked ction 0). Single fr

nthesizer has beenIV.

er processing

adar

cally switchable sn and signal procpattern are depic

c) band airborne sca

e the radiated pu through the rerequency conversn used in the rada

35±0.2 2.5

1.5 2.5

50…500 1…10

3.0 >90 dB 24 dB

50 4-sections

vertical electrical

1 > 25 -13

-53

18-32 DC 70 12 12

slotted cessing cted in

anning

ulse for eceiver sion as ar. The

A prototype of meteorological W band radar has been developed in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine (Vavriv et al, 2002). It utilizes a proprietary magnetron with cold secondary emission cathode. (see Table I). It is featured by: (i) two separate antennas; (ii) a separate downconverter to sample the radiated signal; and (iii) a high power quasi-optical polarization rotators both in Tx and Rx channels. The radar characteristics are summarized in Table V. It should be noticed that the radar in question has been developed several years ago and unfortunately no modifications have been provided till now due to our effort were concentrated on the radars operating within Ka band essentially.

Operating frequency, GHz 94 Peak power (max), kW 4 Pulse width, ns 50 - 400 Pulse repetition frequency, kHz 2.5 - 10 Receiver noise temperature, K 1200 Total dynamical range, dB 70 Polarization HH, VV, HV, VH Cross-polarization isolation, dB -25 Antenna diameter, m 0.5 Antenna beam width, deg 0.45 Total impressed composite wideband noise of Doppler processing @ 5 km, dBc/Hz -47

Sensitivity at 5 km with the integration time of 0.1 sec, dBZ -41

Table V. Parameters of prototype of W band meteorological radar. Therefore a serious improvement of its parameter may be expected due to: (i) utilization of a single antenna due to increase in availability of high power circulators and P-i-N switches for the receiver protection circuitry; (ii) an introduction of digital receiver technique as well as a digital frequency control similarly to the above Ka band radars; (iii) the introduction of a low noise amplifier, which have become available during resent time; (iv) the introduction of a synthesized local oscillator; and (v) the introduction of an advanced magnetron modulator. And at last we would like to provide a brief description W band short pulse transmitter for airport debris radar (Belikov et al, 2002). As known debris are very serious problem to provide enough flight safety. A high resolution as well as a high reliability is a mandatory requirement conceptually for such radars. In the case of magnetron based radar an appropriate resolution may be achieved in a rather simple way, without the utilization of pulse compression. The magnetron with cold-secondary emission cathode used in the transmitter in question provides in addition an extended life time of 10000 hours at least. The parameters of the transmitter are given in Table VI.

Parameter Measure Unit Measured Value Output Frequency(25°C) GHz 95.086 Frequency Temperature Coefficient MHz/°C -1.75 RF Pulse Width (@-3dB, 25°C) ns 17.5 RF Pulse Width (@-6dB, 25°C) ns 20 RF Output Peak Power kW 2

www.intechopen.com


Ta

In de

4.

4.1Leoptheonconspathecoodimsigattto meThis skeyshoin freovme

Fig

RF Pulse JitterPRF Supply VoltagCurrent consuWeight Dimensions

able VI. Parameter

the next sectionscribed briefly.

Magnetron ba

1 General considt us to remind br

perational mode ie spectrum of RF

n the parameters nceptually in locaace of signal parae simplest case ordinates of ampmensions shouldgnal in the space tending the magnprovide a precise

easurement capabhe most evident bsimply to measuy issue to impleould be providedorder to measur

equency is next iverall radar perfoeasured.

g. 5. Typical block

r

ge min umption, max @ 28

rs of W band sho

n some design

ased radars – d

deration riefly that any ms used; (ii) each R oscillation depenof external microating the receivedameters dependin

of non-coherenplitude and time r

be added. In tru of signal param

netron utilization e location of radibilities of the radabut comprehensivre its parameters

ement Doppler pd with correspondre its parametersimportant paramormance. At first

k-diagram of mag

ns kHz

V V A

kg

ort pulse magnetr

approaches used

design approac

agnetron based rRF pulse is charands strongly on aowave circuits. Od signal as respecng of the radar m

nt pulsed radar respectively. For uly coherent rada

meters is known a in the radars reqiated signal in suar. ve method to enss. It is the only wprocessing. Thus ding circuits to sa

s like it is depictemeter, whose mea

t, it determines h

gnetron based rad

19”

on transmitter.

d in the above

ches

radar is featured acterized by an ara shape of modulaOn other hand, ract to the radiated

measurement capathis space is twDoppler radar th

ar systems the exaa priori. Instead, quire introductionuch space and ext

ure exact locationway to get phase i

each magnetronample a small poed in Fig. 5. The

asurement accurahow precisely a

dar.

3 3…30 18…32

14.1 25

”, 5U unit

mentioned rada

as follows: (i) a prbitrary phase; anation voltage as wadar operation co one in a correspoabilities. For examwo dimensional,he phase and freqact location of rathe above peculi

n of specific approtend its dimensio

n of the radiated information, whin based Dopplerortion of radiated e magnetron osciacy affects strong target velocity c

ars are

pulsed nd (iii) well as onsists onding mple in , with

quency adiated iarities oaches

ons, i.e.

signal ich is a r radar signal illation gly the can be

It does not require a great accuracy and may be implemented relatively easily. Practically in the most cases no measurements are required at all due to a specified magnetron frequency deviation does not exceed a portion of percent in the worst case. A different matter is a pulse-to-pulse frequency deviation. This parameter introduces both non-coherent (noise) and regular components (spurs) into Doppler signal processing (see Fig. 3). In part it determines the ability of the radar to resolve targets with different velocities and reflectivity in the same range bin, e.g. clouds in a strong rain or a moving target in presence of a much stronger reflection from a clutter. As it has been exposed above, the magnetron frequency should be measured with accuracy of about 10-7 for a period of several hundred nanoseconds typically or even less, if a higher spatial resolution is required, in order to provide 70 dB spectral dynamical range for Ka band radar and the distance of 5 km. The indicated accuracy is on the edge of contemporary technical capabilities or beyond them, not even to mention the situation inherent to very recent time. Thus at the time being, achieving the maximal possible Doppler performance is the responsibility for the radar circuits, which should ensure as much as possible tight control of magnetron operational parameters – voltage, filament, loading etc, and, finally, its frequency stability. In the nearest future due to a dramatically fast progress in the development of data acquisition and processing hardware we expect that precise measurements of the parameters of the radiated pulse will be a basic method defining radar resolution and instrumentation capabilities. Some promising prospects concerned to this possibility will be discussed later (see Section 4.3.5 ). Below in this section we will try to analyze requirements to high performance magnetron based radar and discuss some methods to meet them.

4.2 Transmitter. 4.2.1 General consideration As mentioned above the modern requirements to the radar performance cannot be met otherwise than designing the magnetron environment to ensure as much as possible stability and safety of its operation. Therefore, the transmitter is probably the most valuable part of either magnetron based high performance radar. Before we will proceed to discuss some design approaches used in the transmitters, let us make a simple calculation in order to give an impression about how precisely its circuits should work. Assume that the aforementioned value of pulse-to-pulse frequency stability 絞血【血 of 10-7 should be provided. The variations of the amplitude of voltage pulse across magnetron should not exceed value given by the following expression:

�撃判捗任濡迩庁寧任如禰 · 岾弟捗捗任濡迩峇 · �鳥 (4)

where 血�鎚� is a magnetron oscillation frequency, �塚�鎮痛 – a magnetron oscillation frequency pushing factor, �鳥 - a dynamical resistance of the magnetron in an operational point, i.e. the slope of its volt-ampere characteristic in this point. Let us take into consideration Ka band magnetron and suggest that the magnetron frequency pushing factor is of 500 kHz/A – a very respectable value, inherent to a highly stable coaxial magnetron rather than any other type, and a dynamical resistance of 300 Ohms – a typical value for devices with 10-100 kW peak power. Then the above expression gives an impressive value of about 2 V, or less than 200 ppm typically, for the required value of pulse-to-pulse amplitude instability of magnetron anode voltage! Note, that the indicated value should be ensured during the

www.intechopen.com


Ta

In de

4.

4.1Leoptheonconspathecoodimsigattto meThis skeyshoin freovme

Fig

RF Pulse JitterPRF Supply VoltagCurrent consuWeight Dimensions

able VI. Parameter

the next sectionscribed briefly.

Magnetron ba

1 General considt us to remind br

perational mode ie spectrum of RF

n the parameters nceptually in locaace of signal parae simplest case ordinates of ampmensions shouldgnal in the space tending the magnprovide a precise

easurement capabhe most evident bsimply to measuy issue to impleould be providedorder to measur

equency is next iverall radar perfoeasured.

g. 5. Typical block

r

ge min umption, max @ 28

rs of W band sho

n some design

ased radars – d

deration riefly that any ms used; (ii) each R oscillation depenof external microating the receivedameters dependin

of non-coherenplitude and time r

be added. In tru of signal param

netron utilization e location of radibilities of the radabut comprehensivre its parameters

ement Doppler pd with correspondre its parametersimportant paramormance. At first

k-diagram of mag

ns kHz

V V A

kg

ort pulse magnetr

approaches used

design approac

agnetron based rRF pulse is charands strongly on aowave circuits. Od signal as respecng of the radar m

nt pulsed radar respectively. For uly coherent rada

meters is known a in the radars reqiated signal in suar. ve method to enss. It is the only wprocessing. Thus ding circuits to sa

s like it is depictemeter, whose mea

t, it determines h

gnetron based rad

19”

on transmitter.

d in the above

ches

radar is featured acterized by an ara shape of modulaOn other hand, ract to the radiated

measurement capathis space is twDoppler radar th

ar systems the exaa priori. Instead, quire introductionuch space and ext

ure exact locationway to get phase i

each magnetronample a small poed in Fig. 5. The

asurement accurahow precisely a

dar.

3 3…30 18…32

14.1 25

”, 5U unit

mentioned rada

as follows: (i) a prbitrary phase; anation voltage as wadar operation co one in a correspoabilities. For examwo dimensional,he phase and freqact location of rathe above peculi

n of specific approtend its dimensio

n of the radiated information, whin based Dopplerortion of radiated e magnetron osciacy affects strong target velocity c

ars are

pulsed nd (iii) well as onsists onding mple in , with

quency adiated iarities oaches

ons, i.e.

signal ich is a r radar signal illation gly the can be

It does not require a great accuracy and may be implemented relatively easily. Practically in the most cases no measurements are required at all due to a specified magnetron frequency deviation does not exceed a portion of percent in the worst case. A different matter is a pulse-to-pulse frequency deviation. This parameter introduces both non-coherent (noise) and regular components (spurs) into Doppler signal processing (see Fig. 3). In part it determines the ability of the radar to resolve targets with different velocities and reflectivity in the same range bin, e.g. clouds in a strong rain or a moving target in presence of a much stronger reflection from a clutter. As it has been exposed above, the magnetron frequency should be measured with accuracy of about 10-7 for a period of several hundred nanoseconds typically or even less, if a higher spatial resolution is required, in order to provide 70 dB spectral dynamical range for Ka band radar and the distance of 5 km. The indicated accuracy is on the edge of contemporary technical capabilities or beyond them, not even to mention the situation inherent to very recent time. Thus at the time being, achieving the maximal possible Doppler performance is the responsibility for the radar circuits, which should ensure as much as possible tight control of magnetron operational parameters – voltage, filament, loading etc, and, finally, its frequency stability. In the nearest future due to a dramatically fast progress in the development of data acquisition and processing hardware we expect that precise measurements of the parameters of the radiated pulse will be a basic method defining radar resolution and instrumentation capabilities. Some promising prospects concerned to this possibility will be discussed later (see Section 4.3.5 ). Below in this section we will try to analyze requirements to high performance magnetron based radar and discuss some methods to meet them.

4.2 Transmitter. 4.2.1 General consideration As mentioned above the modern requirements to the radar performance cannot be met otherwise than designing the magnetron environment to ensure as much as possible stability and safety of its operation. Therefore, the transmitter is probably the most valuable part of either magnetron based high performance radar. Before we will proceed to discuss some design approaches used in the transmitters, let us make a simple calculation in order to give an impression about how precisely its circuits should work. Assume that the aforementioned value of pulse-to-pulse frequency stability 絞血【血 of 10-7 should be provided. The variations of the amplitude of voltage pulse across magnetron should not exceed value given by the following expression:

�撃判捗任濡迩庁寧任如禰 · 岾弟捗捗任濡迩峇 · �鳥 (4)

where 血�鎚� is a magnetron oscillation frequency, �塚�鎮痛 – a magnetron oscillation frequency pushing factor, �鳥 - a dynamical resistance of the magnetron in an operational point, i.e. the slope of its volt-ampere characteristic in this point. Let us take into consideration Ka band magnetron and suggest that the magnetron frequency pushing factor is of 500 kHz/A – a very respectable value, inherent to a highly stable coaxial magnetron rather than any other type, and a dynamical resistance of 300 Ohms – a typical value for devices with 10-100 kW peak power. Then the above expression gives an impressive value of about 2 V, or less than 200 ppm typically, for the required value of pulse-to-pulse amplitude instability of magnetron anode voltage! Note, that the indicated value should be ensured during the

www.intechopen.com


intmaNoor diahigconthama

Fig trathe

4.2ThstamaA (PWinhprochautibautimoassvomucomutifacFoDodePW

terval of data accay vary within thow, when a refere other, it is possibagram of a transmgh voltage powentroller. Let us leat it handles othagnetron operatio

g. 6. Block-diagra

ansmitter with ree magnetron perf

2.2 High voltagehe high voltage pability, i.e. Dopplatter of the highesswitching mode

WM) converter, cherent high efficiovided by such saracteristics of Pilization. Our expsed radars demoilize operation in ode. Such approasists maximizing ltage regulation lultiple to the pmpletely the infilization of a parctor corrector for r information, th

oppler performanveloped accordin

WM converter. Fr

cumulation for Fohe range from tensence point for theble to consider somitter is depicteder supply; (ii) a eave the latter un

her units accordinonal mode as wel

am of magnetron

emote control andformance, thus w

power supply ower supply detler performance ost priority under e power supply,cannot be alternaiency, small dimesupply is lower g

PWM converter mperience to devel

onstrates a benefi either peak curre

ach as well as the both rejection ofloop. Next, it is m

pulse repetition fluence of rippl

rticular pre-regul AC powered syshe line of Ka bnce (see Section ng strictly to the rom our opinion,

ourier processings millisecond up e magnetron tranolutions enablingd in Fig. 6. It incl modulator; (iii)nit beyond a morng the procedurell as provides the

transmitter.

d diagnostics abie would like to o

ermines essentialof whole radar. T the development, based on the uated to produce hensions, and lighgenerally than thmay be improvelop the high voltt of the followingent or close to it m

e usage of a frequf the input voltagmandatory to synfrequency of th

les at PWM opator is preferablytems is virtually

band meteorologi 0) is equipped above recommen, such topology i

g. As usual the dto several portion

nsmitter design is g its consummatioludes the followi a filament powre detailed consides ensuring the m

ilities. Other aboutline their desig

lly the short termThus ensuring itst. utilization of puhigh voltage in m

ht weight. Howevhat of linear regud to an extent atage power suppg rules. At first, Pmixed mode rath

uency compensatege ripples and thenchronize PWM che radar, which erational frequeny. In this respectcompulsorily. ical radar demo with the high ndations. A flybais the most suita

duration of this inns of second. indicated in somon. A simplified ing essential unit

wer supply; and deration, mentiomost optimal an

ove units affect dgn in more detail.

m magnetron freqs maximal stabili

ulse width modumodern systems ver the voltage stulators. On otherallowing its standplies for the magnPWM converter s

her than in pure ved high voltage de overall stabilityonverter at a freq eliminates pracncy. And at las

t, the usage of a

onstrating a veryvoltage power sck topology is usble to the high v

nterval

me way block-ts: (i) a (iv) a n only

nd safe

directly

quency ity is a

ulation due to tability hand, dalone netron should voltage divider y of the quency ctically st, the power

y solid supply sed for voltage

apradis wiswsuphigFighigop

4.2In higincthededesenmichamamomedrashoconpuNomupumawh

Fig

plications with thdar systems or evused. The essent

ithin a wide rangwing across the p

pply voltage. Thgh voltage powerg. 3 there is no rgh voltage powe

perational frequen

2.3 Modulator this section we wgh voltage modcludes circuits to e most cases a velopment. Sinceviation of the pnsitivity. Thus, binimized. Especiaaracterized by a agnetrons requirodulation pulse tean better! An opawn to ensure itsould be taken insiderable thresh

ulse through the otice that at loweuch greater as re

ulse duration andagnetron performhile a pulse repeti

g. 7. Waveforms o

he output powerven airborne DC tial advance of suge of output powprimary windinghe above peculiar

r supply in a maegular spurious

er supply at the ncy of PWM conv

will consider briefulators used in form the pulse wnear-rectangle s

e the magnetron pulse shape fromboth transients aally it is importanrather short widt

res a well contrto facilitate runnipposite situation as appropriately sinto consideratiohold current to pr

magnetron mayer voltages the poespect to anode pd higher pulse r

mance. Thus the ition rate greater

of voltage pulse a

r up to 1 kW and powered radars uch scheme is a

wer as well as ths of the high vo

rities meet perfecagnetron based trcomponents cau harmonics of bo

verter (folded).

fly some issues r high performanwith a definite shshape of RF pufrequency depen

m the rectangulaand the distortiont for the millimeth of the output rollable voltage ing oscillation (Oappears for the tr

short duration. Hon there. It is droduce RF oscilla

y be much longeower of back bompower as indicatrepetition rate thabove issue shou than several kilo

across magnetron

d voltages up to 2 if an appropriatestable operation

he ability to provoltage transformectly actual operatransmitters. As csed by ripples ofoth AC power l

elated to the devnce radars. In ghape across the mulse is a target nds strongly on thar one results inns of flat part o

eter wavelengths pulse. On other hrate during the

Okress, 1961). In thrailing edge. As u

However, not onlydue to the magnation as usual. It er than RF pulsembardment of theted in Fig. 7. Evi

he stronger the auld be always ta

ohertz is required

n and RF envelope

20 kV for AC poe step-up pre-reg with a capacitiv

vide the output ver much greater tional conditions can be easily seenf the output voltline frequency an

velopment of up-tgeneral the modmagnetron termin

under the modhe applied voltagn a drop in the of the pulse shou magnetrons, whihand the most tye leading edge his case faster dousual a less atteny shape of RF envnetrons have a means that the c as depicted in e magnetron cathidently, the shor

above effect affecaken into conside.

e.

owered gulator ve load voltage than a of the n form tage of nd the

to date dulator nals. In dulator ge, any radar uld be ich are

ypes of of the

oes not ntion is velope rather

current Fig. 7.

hode is rter RF cts the eration

www.intechopen.com


intmaNoor diahigconthama

Fig trathe

4.2ThstamaA (PWinhprochautibautimoassvomucomutifacFoDodePW

terval of data accay vary within thow, when a refere other, it is possibagram of a transmgh voltage powentroller. Let us leat it handles othagnetron operatio

g. 6. Block-diagra

ansmitter with ree magnetron perf

2.2 High voltagehe high voltage pability, i.e. Dopplatter of the highesswitching mode

WM) converter, cherent high efficiovided by such saracteristics of Pilization. Our expsed radars demoilize operation in ode. Such approasists maximizing ltage regulation lultiple to the pmpletely the infilization of a parctor corrector for r information, th

oppler performanveloped accordin

WM converter. Fr

cumulation for Fohe range from tensence point for theble to consider somitter is depicteder supply; (ii) a eave the latter un

her units accordinonal mode as wel

am of magnetron

emote control andformance, thus w

power supply ower supply detler performance ost priority under e power supply,cannot be alternaiency, small dimesupply is lower g

PWM converter mperience to devel

onstrates a benefi either peak curre

ach as well as the both rejection ofloop. Next, it is m

pulse repetition fluence of rippl

rticular pre-regul AC powered syshe line of Ka bnce (see Section ng strictly to the rom our opinion,

ourier processings millisecond up e magnetron tranolutions enablingd in Fig. 6. It incl modulator; (iii)nit beyond a morng the procedurell as provides the

transmitter.

d diagnostics abie would like to o

ermines essentialof whole radar. T the development, based on the uated to produce hensions, and lighgenerally than thmay be improvelop the high voltt of the followingent or close to it m

e usage of a frequf the input voltagmandatory to synfrequency of th

les at PWM opator is preferablytems is virtually

band meteorologi 0) is equipped above recommen, such topology i

g. As usual the dto several portion

nsmitter design is g its consummatioludes the followi a filament powre detailed consides ensuring the m

ilities. Other aboutline their desig

lly the short termThus ensuring itst. utilization of puhigh voltage in m

ht weight. Howevhat of linear regud to an extent atage power suppg rules. At first, Pmixed mode rath

uency compensatege ripples and thenchronize PWM che radar, which erational frequeny. In this respectcompulsorily. ical radar demo with the high ndations. A flybais the most suita

duration of this inns of second. indicated in somon. A simplified ing essential unit

wer supply; and deration, mentiomost optimal an

ove units affect dgn in more detail.

m magnetron freqs maximal stabili

ulse width modumodern systems ver the voltage stulators. On otherallowing its standplies for the magnPWM converter s

her than in pure ved high voltage de overall stabilityonverter at a freq eliminates pracncy. And at las

t, the usage of a

onstrating a veryvoltage power sck topology is usble to the high v

nterval

me way block-ts: (i) a (iv) a n only

nd safe

directly

quency ity is a

ulation due to tability hand, dalone netron should voltage divider y of the quency ctically st, the power

y solid supply sed for voltage

apradis wiswsuphigFighigop

4.2In higincthededesenmichamamomedrashoconpuNomupumawh

Fig

plications with thdar systems or evused. The essent

ithin a wide rangwing across the p

pply voltage. Thgh voltage powerg. 3 there is no rgh voltage powe

perational frequen

2.3 Modulator this section we wgh voltage modcludes circuits to e most cases a velopment. Sinceviation of the pnsitivity. Thus, binimized. Especiaaracterized by a agnetrons requirodulation pulse tean better! An opawn to ensure itsould be taken insiderable thresh

ulse through the otice that at loweuch greater as re

ulse duration andagnetron performhile a pulse repeti

g. 7. Waveforms o

he output powerven airborne DC tial advance of suge of output powprimary windinghe above peculiar

r supply in a maegular spurious

er supply at the ncy of PWM conv

will consider briefulators used in form the pulse wnear-rectangle s

e the magnetron pulse shape fromboth transients aally it is importanrather short widt

res a well contrto facilitate runnipposite situation as appropriately sinto consideratiohold current to pr

magnetron mayer voltages the poespect to anode pd higher pulse r

mance. Thus the ition rate greater

of voltage pulse a

r up to 1 kW and powered radars uch scheme is a

wer as well as ths of the high vo

rities meet perfecagnetron based trcomponents cau harmonics of bo

verter (folded).

fly some issues r high performanwith a definite shshape of RF pufrequency depen

m the rectangulaand the distortiont for the millimeth of the output rollable voltage ing oscillation (Oappears for the tr

short duration. Hon there. It is droduce RF oscilla

y be much longeower of back bompower as indicatrepetition rate thabove issue shou than several kilo

across magnetron

d voltages up to 2 if an appropriatestable operation

he ability to provoltage transformectly actual operatransmitters. As csed by ripples ofoth AC power l

elated to the devnce radars. In ghape across the mulse is a target nds strongly on thar one results inns of flat part o

eter wavelengths pulse. On other hrate during the

Okress, 1961). In thrailing edge. As u

However, not onlydue to the magnation as usual. It er than RF pulsembardment of theted in Fig. 7. Evi

he stronger the auld be always ta

ohertz is required

n and RF envelope

20 kV for AC poe step-up pre-reg with a capacitiv

vide the output ver much greater tional conditions can be easily seenf the output voltline frequency an

velopment of up-tgeneral the modmagnetron termin

under the modhe applied voltagn a drop in the of the pulse shou magnetrons, whihand the most tye leading edge his case faster dousual a less atteny shape of RF envnetrons have a means that the c as depicted in e magnetron cathidently, the shor

above effect affecaken into conside.

e.

owered gulator ve load voltage than a of the n form tage of nd the

to date dulator nals. In dulator ge, any radar uld be ich are

ypes of of the

oes not ntion is velope rather

current Fig. 7.

hode is rter RF cts the eration

www.intechopen.com


NodepatheoumoutiA linbo

Fig Thconalscurin uti(iiibreovproHothepudusignosenmatrautito all radmaAnwifro

ow, when the esssign approaches rts: (i) energy stoe magnetron wit

utput and the maodulator types beilized practically simplified block-

ne arranged as a pth to store energy

g. 8. Block diagram

he utilization of ntrolled devices lso thyristors or rrents and voltaggeneral; and do

ilization of the stei) practically comeakdown due to

verall complexity obably the lowesowever there are e usage of a lum

ulse. Next, it is pruration. Further, tgnificant distortioonlinearity of mansitivity from thatching the impeailing edge of thilizing modulatorbe used in the hi mentioned disaddar performanceagnetron based synother type of hiith partial discharom the line mod

sential requireme to meet them canorage; (ii) a switcthin definite timagnetron; and (vieing in use, a lineexclusively in ma-diagram of a linpiece of cable or ay and to form the

m of line modula

line modulators like a hard tubes thyratrons, whic

ges (MOSFET vs. o not require a ep-up transformemplete cancellatithe total energy s and total cost ot as compared wi a number of serio

mped element delaractically impossithe inherent utilions of the modugnetron volt-amp

he voltage rate aedances of the d

he modulation purs equipped withgh performance rdvantages of the e only and causystems. igh voltage modurge (Sivan, 1994)

dulator is only a

ents to the modun be discussed. Ech or switches, w

me intervals onlyi) protection ande modulator and agnetron transmie modulator is d

assembled from lue trailing edge of t

ator.

provides the fo or MOSFETs maych are rated fun thyristors compaprecise shape o

er to match the imion of the possibstored in the delaof a transmitter ith other types. ous disadvantageay line leads us tible to implemenization of the tra

ulator output pupere characteristi

across the magnedelay line and thulse. Both nume

h a transformer prradars. On other line modulators se no problems

ulator used to dr. The essential d

a small part of e

ulator circuits havEither modulator which provides ay; (iii) circuits to d decoupling circ

a modulator wititters (Sivan, 1994

depicted schematiump inductors anthe pulse across m

ollowing advantay be used in the hndamentally to arison); demonstrof triggering pulmpedances of delble magnetron day line is limited.equipped with t

es of such type ofto oscillations apt smooth regulati

ansformer resultsulse especially duic as well as posetron (Okress, 19he magnetron drical simulation revent us to reco hands it should b are important toin a variety of

rive magnetrons ifference of suchenergy accumula

ve been outlined, contains the foll

applying voltage match the moduits. From a varith partial dischar4). ically in Fig. 8. And capacitors is umagnetron.

ages: (i) not onlyhigh voltage switwork at much rate a greater effise; (ii) the manay line and magn

damage in the c. All above resultthe line modulat

f modulator. The ppear at flat part ing of the output

s in the introductue to there is a ssible appearance961). It makes during the leadinand the experie

ommend such appbe accepted that a

o achieve the max simple and low

is so called modh type of the modated in an appro

, some lowing across

dulator iety of rge are

A delay used as

y fully tch but higher iciency datory netron; case of t in the tor are

first is of the t pulse tion of strong

e of its ifficult

ng and ence of proach almost ximum w cost

dulator dulator opriate

stomautipuA badisfreis biamoa sto

Fig A podiscomordreschoa pmocomgenpacompawion200A in poconadde

orage is used to fagnetron modulailization of match

ulse than the line mvariety of variansic configurationsadvantages provequently than otha great advantagas power supplieodulator design. Tstack of transistordrive each transi

a) g. 9. Schemes of th

resistor or chokeower supply durstortions of magmponent with frder to decrease sistor should be oke is used insteapossibility to prododulator supply vmponent, the outnerally, there arerameters taking mponent is bulkrts of the modula

ith inductive deconly if retrieving of02)), such schemesimplified block-Fig 9.b The only

ower supply and nceptually in thisditional resistor picted in Fig. 9.b

form the output ators exclusively. hing circuitry conmodulator. nts can be used tns are depicted vided by each o

her configurationsge for whichever s for the modulatThis issue is not srs connected in sestor independent

he modulator wit

e may be used toing interval of p

gnetron pulse durequency dependduration of the chosen small enad the resistor, thduce the output voltage. Howevetput pulse distorte big constructivinto account it i

ky and characteriator. Thus we wooupling in modef very short pulsee may appear to b-diagram of the my capacitor may b storage capacitos case. Neverthelto decrease the d

b by a dash line. I

pulse. A capacit Due to such typnceptually it prov

to build the mod in Fig. 9. Let of them. The firs. It is due to the hard tube basedtor tube are grouso important for eries. Actually, a gtly on either the h

b) th partial dischar

o decouple the spulse formation. uring its front adant impedance trailing edge of ough, which resuhe efficiency is bepulse with amplir due to the pulstions are higher in

ve difficulties to dis under high puized by a ratherould not like to r

ern high performaes at a high repetbe the most optim

modulator with a be used both as tr of the modulatless, as mentioneduration of the oIt is difficult evid

tor is used as thee of the modulat

vides a much bet

dulator with part us consider brrst scheme is trhigh voltage swi

d modulator. In thunded also, simpl

a solid state switgalvanic decouplhigh voltage switc

rge.

switch from the o In the case of rand flat part aris in the pulse the output puls

ults in lower moetter fundamentaitude, which is he formation netwn the case of the cdesign a charginulsed voltage con large weight as

recommend the uance magnetron tition rate is requ

mal choice. floating high volthe output capacitor. No decouplined above if a highutput pulse traili

dently to use a ha

e energy storage tor does not requter shape of the o

tial discharge. Soriefly advantageaditionally useditch is grounded, his case a filamenlifying considerabtch, arranged usuling is required anch is grounded o

c)

output of high vresistor utilizatiore minimal due formation netwoe the resistance

odulator efficiencally. In addition thhigher than the vawork includes a re

choke utilizationng choke with reqndition. As usuas compared withutilization of modbased radars. Pro

uired (see (Belikov

ltage switch is deitor of the high vng circuits are reqh PRF is suggesting edge is requiard tube in such

in the uire the output

ome of es and d more which nt and bly the

ually as nyway r not.

voltage on, the

to no ork. In of the

cy. If a here is alue of eactive . Next, quired

al such h other dulator obably v et al,

epicted voltage quired ted, an ired as circuit

www.intechopen.com


NodepatheoumoutiA linbo

Fig Thconalscurin uti(iiibreovproHothepudusignosenmatrautito all radmaAnwifro

ow, when the esssign approaches rts: (i) energy stoe magnetron wit

utput and the maodulator types beilized practically simplified block-

ne arranged as a pth to store energy

g. 8. Block diagram

he utilization of ntrolled devices lso thyristors or rrents and voltaggeneral; and do

ilization of the stei) practically comeakdown due to

verall complexity obably the lowesowever there are e usage of a lum

ulse. Next, it is pruration. Further, tgnificant distortioonlinearity of mansitivity from thatching the impeailing edge of thilizing modulatorbe used in the hi mentioned disaddar performanceagnetron based synother type of hiith partial discharom the line mod

sential requireme to meet them canorage; (ii) a switcthin definite timagnetron; and (vieing in use, a lineexclusively in ma-diagram of a linpiece of cable or ay and to form the

m of line modula

line modulators like a hard tubes thyratrons, whic

ges (MOSFET vs. o not require a ep-up transformemplete cancellatithe total energy s and total cost ot as compared wi a number of serio

mped element delaractically impossithe inherent utilions of the modugnetron volt-amp

he voltage rate aedances of the d

he modulation purs equipped withgh performance rdvantages of the e only and causystems. igh voltage modurge (Sivan, 1994)

dulator is only a

ents to the modun be discussed. Ech or switches, w

me intervals onlyi) protection ande modulator and agnetron transmie modulator is d

assembled from lue trailing edge of t

ator.

provides the fo or MOSFETs maych are rated fun thyristors compaprecise shape o

er to match the imion of the possibstored in the delaof a transmitter ith other types. ous disadvantageay line leads us tible to implemenization of the tra

ulator output pupere characteristi

across the magnedelay line and thulse. Both nume

h a transformer prradars. On other line modulators se no problems

ulator used to dr. The essential d

a small part of e

ulator circuits havEither modulator which provides ay; (iii) circuits to d decoupling circ

a modulator wititters (Sivan, 1994

depicted schematiump inductors anthe pulse across m

ollowing advantay be used in the hndamentally to arison); demonstrof triggering pulmpedances of delble magnetron day line is limited.equipped with t

es of such type ofto oscillations apt smooth regulati

ansformer resultsulse especially duic as well as posetron (Okress, 19he magnetron drical simulation revent us to reco hands it should b are important toin a variety of

rive magnetrons ifference of suchenergy accumula

ve been outlined, contains the foll

applying voltage match the moduits. From a varith partial dischar4). ically in Fig. 8. And capacitors is umagnetron.

ages: (i) not onlyhigh voltage switwork at much rate a greater effise; (ii) the manay line and magn

damage in the c. All above resultthe line modulat

f modulator. The ppear at flat part ing of the output

s in the introductue to there is a ssible appearance961). It makes during the leadinand the experie

ommend such appbe accepted that a

o achieve the max simple and low

is so called modh type of the modated in an appro

, some lowing across

dulator iety of rge are

A delay used as

y fully tch but higher iciency datory netron; case of t in the tor are

first is of the t pulse tion of strong

e of its ifficult

ng and ence of proach almost ximum w cost

dulator dulator opriate

stomautipuA badisfreis biamoa sto

Fig A podiscomordreschoa pmocomgenpacompawion200A in poconadde

orage is used to fagnetron modulailization of match

ulse than the line mvariety of variansic configurationsadvantages provequently than otha great advantagas power supplieodulator design. Tstack of transistordrive each transi

a) g. 9. Schemes of th

resistor or chokeower supply durstortions of magmponent with frder to decrease sistor should be oke is used insteapossibility to prododulator supply vmponent, the outnerally, there arerameters taking mponent is bulkrts of the modula

ith inductive deconly if retrieving of02)), such schemesimplified block-Fig 9.b The only

ower supply and nceptually in thisditional resistor picted in Fig. 9.b

form the output ators exclusively. hing circuitry conmodulator. nts can be used tns are depicted vided by each o

her configurationsge for whichever s for the modulatThis issue is not srs connected in sestor independent

he modulator wit

e may be used toing interval of p

gnetron pulse durequency dependduration of the chosen small enad the resistor, thduce the output voltage. Howevetput pulse distorte big constructivinto account it i

ky and characteriator. Thus we wooupling in modef very short pulsee may appear to b-diagram of the my capacitor may b storage capacitos case. Neverthelto decrease the d

b by a dash line. I

pulse. A capacit Due to such typnceptually it prov

to build the mod in Fig. 9. Let of them. The firs. It is due to the hard tube basedtor tube are grouso important for eries. Actually, a gtly on either the h

b) th partial dischar

o decouple the spulse formation. uring its front adant impedance trailing edge of ough, which resuhe efficiency is bepulse with amplir due to the pulstions are higher in

ve difficulties to dis under high puized by a ratherould not like to r

ern high performaes at a high repetbe the most optim

modulator with a be used both as tr of the modulatless, as mentioneduration of the oIt is difficult evid

tor is used as thee of the modulat

vides a much bet

dulator with part us consider brrst scheme is trhigh voltage swi

d modulator. In thunded also, simpl

a solid state switgalvanic decouplhigh voltage switc

rge.

switch from the o In the case of rand flat part aris in the pulse the output puls

ults in lower moetter fundamentaitude, which is he formation netwn the case of the cdesign a charginulsed voltage con large weight as

recommend the uance magnetron tition rate is requ

mal choice. floating high volthe output capacitor. No decouplined above if a highutput pulse traili

dently to use a ha

e energy storage tor does not requter shape of the o

tial discharge. Soriefly advantageaditionally useditch is grounded, his case a filamenlifying considerabtch, arranged usuling is required anch is grounded o

c)

output of high vresistor utilizatiore minimal due formation netwoe the resistance

odulator efficiencally. In addition thhigher than the vawork includes a re

choke utilizationng choke with reqndition. As usuas compared withutilization of modbased radars. Pro

uired (see (Belikov

ltage switch is deitor of the high vng circuits are reqh PRF is suggesting edge is requiard tube in such

in the uire the output

ome of es and d more which nt and bly the

ually as nyway r not.

voltage on, the

to no ork. In of the

cy. If a here is alue of eactive . Next, quired

al such h other dulator obably v et al,

epicted voltage quired ted, an ired as circuit

www.intechopen.com


configuration. Instead, from our opinion, it is the most preferable scheme to utilize a solid state high voltage switch. The push-pull scheme of the modulator is depicted in Fig. 9, c. It provides the tightest control of the output pulse shape as well as the highest energy efficiency among the schemes discussed before. The expense for that is much more complexity in design as compared to previous solutions. It is one of the reasons why the modulators utilizing such approach may be found in a very limited number of radars despite its evident advantages. Let us now discuss some issues concerned to the selection of an appropriate electronic device to build the high voltage switch. According to modern tendencies solid state devices should be considered as first choice while designing any new electronic system. Two types of solid state devices, namely, MOSFETS, and IGBT may be used in magnetron modulators. Despite IGBT are superior generally to MOSFETs as respect to both maximal rated voltage and current, as well as efficiency provided, they are characterized by a lower switching speed and an attitude to a second-induced breakdown, which confines the overall reliability of the modulator. Thus, MOSFETs remain the only choice among solid state devices to use in the magnetron modulators intended for high performance millimeter wavelength radars featured by rather short operational pulse width. Hard tubes were historically the first devices used to build high voltage modulators in radars. They remain to be utilized widely until now despite a serious competition from solid state devices. A considerably greater robustness should be indicated as an essential reason. High voltage circuitry have a strong attitude to appearance of various local breakdowns, leakages etc, which are difficult to control and prevent especially for long-term unattended radar operation. Such phenomena stress the modulator parts greatly. Energy to destroy a hard tube is on orders higher than that for any solid state device. Next, the only tube can be used practically always to build any modulator. Instead the limitations for currently available powerful MOSFETs in the maximal rated voltage, makes inevitable utilizing a stack arranged from many transistors connected in series and, possibly, parallel in order to achieve the modulator parameter being enough to drive the most magnetrons. Utilization of pulse transformer allows in principle to minimize number of the transistors used but, as mentioned above, causes considerable pulse distortions. Certainly there is a well-known disadvantage of hard tubes, namely, a limited life time. We dare to claim that it may be considered as almost virtual at the time being. The situation is very similar to that mentioned above for the magnetrons. A current state of cathode manufacturing as well general state of vacuum technique makes expected lifetime of the modulator hard tubes of several ten thousand hours very realistic. These expectations were ascertained completely by our experience of utilization of hard tube based modulators in the line of meteorological radars (see Section 3). The above consideration allows us to make a conclusion that despite of a strong competition from solid state devices, partially MOSFETs, hard tubes are keeping their positions under development of modern magnetron based radars. The only issue may prevent using them in the modulators, namely, their commercial availability and assortment, which decreases actually generally at the time being due to, essentially, a shortage in demand from non-radar application. Certainly we do not mean a fantastic breakthrough in the development of high voltage, high power semiconductor devices, which makes all hard tubes obsolete at one bout! It should be noticed that the modulator may operates as either voltage or current source. The latter is conceptually better for any cross-field vacuum tube (Sivan, 1994). On other

hand the introduction of the current mode in the modulator makes its design more complex especially in the case if a short pulse length is required. Our experience demonstrates that providing an appropriate stability of the high voltage power supply it is possible to achieve a great magnetron performance even if the much simpler voltage mode is utilized in the modulator. Nevertheless, the development of a modulator, operating in the current mode, remains the greatest challenge a designer faces from our opinion.

4.2.4 Magnetron filament and protection circuits As mentioned in Section 0, keeping an adequate condition of the cathode surface is the most important issue to prolong the magnetron lifetime. It depends on the following factors: (i) cathode temperature; (iii) vacuum condition inside the magnetron; (iii) electron back bombardment. The cathode temperature depends both of a filament power applied and the power dissipated on the cathode due to its back bombardment. This temperature should be kept within a rather narrow interval of several tens degrees typically. Thus, the filament power supply should ensure a very tight control of the magnetron filament power as well as provide a dedicated procedure to regulate it depending on the parameters of the magnetron operational mode. In the developed radars (see Section 0) the following proven principle is used. The filament power supply comprises of two parts, low and high side ones correspondingly. The low side part is simply PWM inverter equipped with either analog or digital controller. The high side includes a high voltage decoupling transformer, a rectifier, and a dedicated controller. It should be noticed that we consider that DC filament voltage should be used to supply magnetron filament due to in this case a possible alternating of the magnetron frequency is canceled. The controller is used to measure both filament voltage and current and to transfer the corresponding data to the low side in a digital form. An optical link is used for such communication. The above approach provides accurate and independent measurement and control of the magnetron filament parameters. The vacuum conditions inside the magnetrons depend not only on quality of manufacturing routine and materials it is made of. Electrical breakdowns affect them strongly. In general, it is considered that the magnetrons demonstrate a rather strong attitude to the development of breakdowns. They may cause in addition direct magnetron damage if no current limiting is provided by the modulator circuitry. It is important for the millimeter wavelengths magnetrons especially due to such magnetrons are characterized by a rather delicate structure of cavities. Hard tubes provide inherently such current limitation, which can be regulated moreover. MOSFETs are featured similarly, but it is difficult to regulate a limit value due to the high voltage switch consists always of several devices connected in series. As a result, this limit is defined by the maximal rated current of the transistors used. Certainly, there is a contradiction between the necessity both to limit the output current and to ensure a minimal settling time of the output voltage of the modulator. Our experience demonstrates that there is a simple way to prevent possible worsening of the magnetron parameters due to breakdowns without a noticeable degradation in the shape of its output pulse. Namely ferrite beads should be connected in series with the magnetron. This effect may be explained as follows. Breakdown in the magnetron appears usually in two stages. The first stage caused by field emission from a tip located somewhere inside magnetron. This stage runs very fast, typically during several nanoseconds. It cause negligible damages of the magnetron internal parts, but initiates developing the next stage

www.intechopen.com


configuration. Instead, from our opinion, it is the most preferable scheme to utilize a solid state high voltage switch. The push-pull scheme of the modulator is depicted in Fig. 9, c. It provides the tightest control of the output pulse shape as well as the highest energy efficiency among the schemes discussed before. The expense for that is much more complexity in design as compared to previous solutions. It is one of the reasons why the modulators utilizing such approach may be found in a very limited number of radars despite its evident advantages. Let us now discuss some issues concerned to the selection of an appropriate electronic device to build the high voltage switch. According to modern tendencies solid state devices should be considered as first choice while designing any new electronic system. Two types of solid state devices, namely, MOSFETS, and IGBT may be used in magnetron modulators. Despite IGBT are superior generally to MOSFETs as respect to both maximal rated voltage and current, as well as efficiency provided, they are characterized by a lower switching speed and an attitude to a second-induced breakdown, which confines the overall reliability of the modulator. Thus, MOSFETs remain the only choice among solid state devices to use in the magnetron modulators intended for high performance millimeter wavelength radars featured by rather short operational pulse width. Hard tubes were historically the first devices used to build high voltage modulators in radars. They remain to be utilized widely until now despite a serious competition from solid state devices. A considerably greater robustness should be indicated as an essential reason. High voltage circuitry have a strong attitude to appearance of various local breakdowns, leakages etc, which are difficult to control and prevent especially for long-term unattended radar operation. Such phenomena stress the modulator parts greatly. Energy to destroy a hard tube is on orders higher than that for any solid state device. Next, the only tube can be used practically always to build any modulator. Instead the limitations for currently available powerful MOSFETs in the maximal rated voltage, makes inevitable utilizing a stack arranged from many transistors connected in series and, possibly, parallel in order to achieve the modulator parameter being enough to drive the most magnetrons. Utilization of pulse transformer allows in principle to minimize number of the transistors used but, as mentioned above, causes considerable pulse distortions. Certainly there is a well-known disadvantage of hard tubes, namely, a limited life time. We dare to claim that it may be considered as almost virtual at the time being. The situation is very similar to that mentioned above for the magnetrons. A current state of cathode manufacturing as well general state of vacuum technique makes expected lifetime of the modulator hard tubes of several ten thousand hours very realistic. These expectations were ascertained completely by our experience of utilization of hard tube based modulators in the line of meteorological radars (see Section 3). The above consideration allows us to make a conclusion that despite of a strong competition from solid state devices, partially MOSFETs, hard tubes are keeping their positions under development of modern magnetron based radars. The only issue may prevent using them in the modulators, namely, their commercial availability and assortment, which decreases actually generally at the time being due to, essentially, a shortage in demand from non-radar application. Certainly we do not mean a fantastic breakthrough in the development of high voltage, high power semiconductor devices, which makes all hard tubes obsolete at one bout! It should be noticed that the modulator may operates as either voltage or current source. The latter is conceptually better for any cross-field vacuum tube (Sivan, 1994). On other

hand the introduction of the current mode in the modulator makes its design more complex especially in the case if a short pulse length is required. Our experience demonstrates that providing an appropriate stability of the high voltage power supply it is possible to achieve a great magnetron performance even if the much simpler voltage mode is utilized in the modulator. Nevertheless, the development of a modulator, operating in the current mode, remains the greatest challenge a designer faces from our opinion.

4.2.4 Magnetron filament and protection circuits As mentioned in Section 0, keeping an adequate condition of the cathode surface is the most important issue to prolong the magnetron lifetime. It depends on the following factors: (i) cathode temperature; (iii) vacuum condition inside the magnetron; (iii) electron back bombardment. The cathode temperature depends both of a filament power applied and the power dissipated on the cathode due to its back bombardment. This temperature should be kept within a rather narrow interval of several tens degrees typically. Thus, the filament power supply should ensure a very tight control of the magnetron filament power as well as provide a dedicated procedure to regulate it depending on the parameters of the magnetron operational mode. In the developed radars (see Section 0) the following proven principle is used. The filament power supply comprises of two parts, low and high side ones correspondingly. The low side part is simply PWM inverter equipped with either analog or digital controller. The high side includes a high voltage decoupling transformer, a rectifier, and a dedicated controller. It should be noticed that we consider that DC filament voltage should be used to supply magnetron filament due to in this case a possible alternating of the magnetron frequency is canceled. The controller is used to measure both filament voltage and current and to transfer the corresponding data to the low side in a digital form. An optical link is used for such communication. The above approach provides accurate and independent measurement and control of the magnetron filament parameters. The vacuum conditions inside the magnetrons depend not only on quality of manufacturing routine and materials it is made of. Electrical breakdowns affect them strongly. In general, it is considered that the magnetrons demonstrate a rather strong attitude to the development of breakdowns. They may cause in addition direct magnetron damage if no current limiting is provided by the modulator circuitry. It is important for the millimeter wavelengths magnetrons especially due to such magnetrons are characterized by a rather delicate structure of cavities. Hard tubes provide inherently such current limitation, which can be regulated moreover. MOSFETs are featured similarly, but it is difficult to regulate a limit value due to the high voltage switch consists always of several devices connected in series. As a result, this limit is defined by the maximal rated current of the transistors used. Certainly, there is a contradiction between the necessity both to limit the output current and to ensure a minimal settling time of the output voltage of the modulator. Our experience demonstrates that there is a simple way to prevent possible worsening of the magnetron parameters due to breakdowns without a noticeable degradation in the shape of its output pulse. Namely ferrite beads should be connected in series with the magnetron. This effect may be explained as follows. Breakdown in the magnetron appears usually in two stages. The first stage caused by field emission from a tip located somewhere inside magnetron. This stage runs very fast, typically during several nanoseconds. It cause negligible damages of the magnetron internal parts, but initiates developing the next stage

www.intechopen.com


brought by a local overheating, cooper sputtering, and further arcing. This stage requires much longer time to run but can cause significant degradation of the magnetron performance. The ferrite beads demonstrate rather high impedance under transient conditions. Thus, it leads to fast decrease in the voltage across magnetron during the first stage of the breakdown and prevents developing of the second stage. Moreover the magnetron keeps producing RF oscillations as usual. There are some more benefits from the utilization of ferrite beads. They restrict the change rate of magnetron current resulting in a great reduction of electromagnetic interferences and an increase to overall stability of the transmitter operation. In addition such beads may improve the output pulse shape weakening the influence of stray circuit inductance and capacitance on the pulse front formation. It should be noticed that the usage of ferrite beads is practically mandatory if MOSFET based high voltage switch is used in the modulator. Such devices are characterized by rather low rated �� capability. According to a numerical simulation and experimental investigations, an excess in this value is an essential reason of transistors damage during transients in the magnetron. At last, it should be noticed that there are evidences that influence of the cathode back bombardment is not limited by thermal effects only as mentioned above. Change in the cathode surface condition has been fixed even after single pulses. Probably it is due to a very high value of instantaneous power of the back bombardment, which is about 10 % of anode power at a nominal anode voltages and much more – up to 30 % - at lower voltages. In order to minimize effect of back bombardment it is necessity to keep the duration of the modulation pulse trailing edge as shorter as possible (see Section 0)

4.3 Receiver 4.3.1 General consideration The principles to develop receivers for the magnetron based radars are the same as for any other systems. Nevertheless, we would like to discuss below some design approaches, which have proven their efficiency in a number of radars developed in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine. A simplified block-diagram of a full-featured double channel receiver used in the most modern modifications of Ka band magnetron based meteorological radars (see Section 0) is depicted in Fig. 12. It may be considered as typical and reflecting the essential design approaches. At first, double frequency conversion is used in the receiver. It makes the receiver design more complex apparently, but simplifies greatly the design of local oscillators as well as allows utilizing a simple filtration to reject image frequency. The local oscillator(s) should meet the following requirements: (i) cover wide enough frequency range – typical values of possible frequency variation are 500 MHz for Ka band and 700…1000 MHz for W and G band magnetrons respectively; (ii) provide frequency tuning with a relatively small step; and (iii) ensure a low phase noise with a fast roll-off beyond the band of �� 2� �� 2� �, which corresponds to the unambiguous frequency range of Doppler processing.

Fig ThprofixoscIf qfreNepueviproInsillufredis(ii)phpicNesysdirmu

g. 10. Block-diagr

he latter is requiocessing. In the c

xed frequency, whcillator with freqquality of Doppleequency conversioext point we wouulse to the data ident way is to uovides the highestead a signal, leustrated by Fig. equency only. Thsadvantages: (i) a) as usual the rece

hase instability. Sck a sample of theevertheless it wastems, which userect comparison uch more design

ram of receiver fo

ired to avoid ancase of double frehich makes its de

quency multiplierer processing proon simplifies its d

uld like to discus acquisition uniuse a separate doest possible perfoeaked through th

11. In this case his way is certaian additional phaeiver circuitry op

Several radars hae transmitted pulas worse on seve a separate down

is hardly possibdetails than the m

or magnetron base

n increase in theequency conversisign relatively sim

r provides enougovided by a radardesign significants is a methods tot to implement

ownconverting chormance but req

he receiver protec newly introduceinly much simplase jitter may be iperates in a deep ave been developlse up. They demveral decibels asnconverting chanble due to the symethod of sampli

ed Doppler radar

e noise floor of ion the first local mple. As usual a

gh performance ler is not an issue, tly (Volkov et al, o couple a sample

coherence-on-rehannel as depicte

quires additional ction circuitry, med circuits operaler but characterintroduced by an saturation, whichped by utilizing

monstrated a solids compared wit

nnel. It should beystems in considing transmitted p

r.

coherence-on-re oscillator operat dielectrically stabevel for the mostthe utilization of 2007) e of the transmitt

eceiver processined in Fig. 10. Thi hardware to be

may be used in thate at the intermrized by the folln antenna environh may cause addthe above appro

d Doppler performth the performane noticed actuallyderation distingu

pulse.

eceiver tes at a bilized t cases. single

ted RF ng. An is way

e used. he way

mediate lowing nment;

ditional oach to mance. nce of

y that a uish in

www.intechopen.com


brought by a local overheating, cooper sputtering, and further arcing. This stage requires much longer time to run but can cause significant degradation of the magnetron performance. The ferrite beads demonstrate rather high impedance under transient conditions. Thus, it leads to fast decrease in the voltage across magnetron during the first stage of the breakdown and prevents developing of the second stage. Moreover the magnetron keeps producing RF oscillations as usual. There are some more benefits from the utilization of ferrite beads. They restrict the change rate of magnetron current resulting in a great reduction of electromagnetic interferences and an increase to overall stability of the transmitter operation. In addition such beads may improve the output pulse shape weakening the influence of stray circuit inductance and capacitance on the pulse front formation. It should be noticed that the usage of ferrite beads is practically mandatory if MOSFET based high voltage switch is used in the modulator. Such devices are characterized by rather low rated �� capability. According to a numerical simulation and experimental investigations, an excess in this value is an essential reason of transistors damage during transients in the magnetron. At last, it should be noticed that there are evidences that influence of the cathode back bombardment is not limited by thermal effects only as mentioned above. Change in the cathode surface condition has been fixed even after single pulses. Probably it is due to a very high value of instantaneous power of the back bombardment, which is about 10 % of anode power at a nominal anode voltages and much more – up to 30 % - at lower voltages. In order to minimize effect of back bombardment it is necessity to keep the duration of the modulation pulse trailing edge as shorter as possible (see Section 0)

4.3 Receiver 4.3.1 General consideration The principles to develop receivers for the magnetron based radars are the same as for any other systems. Nevertheless, we would like to discuss below some design approaches, which have proven their efficiency in a number of radars developed in the Institute of Radio Astronomy of National Academy of Sciences of Ukraine. A simplified block-diagram of a full-featured double channel receiver used in the most modern modifications of Ka band magnetron based meteorological radars (see Section 0) is depicted in Fig. 12. It may be considered as typical and reflecting the essential design approaches. At first, double frequency conversion is used in the receiver. It makes the receiver design more complex apparently, but simplifies greatly the design of local oscillators as well as allows utilizing a simple filtration to reject image frequency. The local oscillator(s) should meet the following requirements: (i) cover wide enough frequency range – typical values of possible frequency variation are 500 MHz for Ka band and 700…1000 MHz for W and G band magnetrons respectively; (ii) provide frequency tuning with a relatively small step; and (iii) ensure a low phase noise with a fast roll-off beyond the band of �� 2� �� 2� �, which corresponds to the unambiguous frequency range of Doppler processing.

Fig ThprofixoscIf qfreNepueviproInsillufredis(ii)phpicNesysdirmu

g. 10. Block-diagr

he latter is requiocessing. In the c

xed frequency, whcillator with freqquality of Doppleequency conversioext point we wouulse to the data ident way is to uovides the highestead a signal, leustrated by Fig. equency only. Thsadvantages: (i) a) as usual the rece

hase instability. Sck a sample of theevertheless it wastems, which userect comparison uch more design

ram of receiver fo

ired to avoid ancase of double frehich makes its de

quency multiplierer processing proon simplifies its d

uld like to discus acquisition uniuse a separate doest possible perfoeaked through th

11. In this case his way is certaian additional phaeiver circuitry op

Several radars hae transmitted pulas worse on seve a separate down

is hardly possibdetails than the m

or magnetron base

n increase in theequency conversisign relatively sim

r provides enougovided by a radardesign significants is a methods tot to implement

ownconverting chormance but req

he receiver protec newly introduceinly much simplase jitter may be iperates in a deep ave been developlse up. They demveral decibels asnconverting chanble due to the symethod of sampli

ed Doppler radar

e noise floor of ion the first local mple. As usual a

gh performance ler is not an issue, tly (Volkov et al, o couple a sample

coherence-on-rehannel as depicte

quires additional ction circuitry, med circuits operaler but characterintroduced by an saturation, whichped by utilizing

monstrated a solids compared wit

nnel. It should beystems in considing transmitted p

r.

coherence-on-re oscillator operat dielectrically stabevel for the mostthe utilization of 2007) e of the transmitt

eceiver processined in Fig. 10. Thi hardware to be

may be used in thate at the intermrized by the folln antenna environh may cause addthe above appro

d Doppler performth the performane noticed actuallyderation distingu

pulse.

eceiver tes at a bilized t cases. single

ted RF ng. An is way

e used. he way

mediate lowing nment;

ditional oach to mance. nce of

y that a uish in

www.intechopen.com


Fig

4.3Aswhstacomproreqsimdulimbe recSois duspidetraW twThaplevconbaevicirpro

g. 11. Utilization o

3.2 Receiver pros known the maghereas maximal rage of the radar mbination of circotection. It mayquired to controlmilar approach foue to a lack of suimiters allow so-ca much less at thcovery time of thme types of magncharacterized of

uration, which mikes in the passivtector in order t

ade-off between th band high peak

wo separate antennhe switches basedproaches should

vel of both peaknfiguration. It mased switch as it iidently in the sw

rculator used, whovided by the dio

of leakage signal

otection circuitrygnetron is characrated input pow receiver, is of m

culator and limite be designed byl them. It increasor the radars operitable limiting dioalled spikes, whiche front of RF phe diode. Evidennetrons, partially

ften by a hard smay be of 2-5 nse

ve limiters provido bias the limitehe losses introdu

k power inherentnas for the transmd on P-i-N diode be used to meetk and average pakes its design rais depicted in Figwitch. Attenuati

hereas insertion loode.

to implement coh

y cterized by a rather of low noise amilliwatt level. W

er is the standard y using only pases reliability conrating within millodes accepting kch mean that the pulse than durinntly the shorter Ry the magnetronsself-excitation resec only. There arding, for example

er diode during Ruced by the biasint to the magnetromitter and receivees provide better t above requirempower a switchather complicated

g. 12, a. A reflection in such confosses is determin

herence-on-receiv

her high both peaamplifiers (LNAWithin microwav de facto solutionassive circuits. Tnsiderably. Howelimeter wavelengilowatt level pea attenuation prov

ng its flat part dRF pulse front th with cold secondsulting in a veryre methods to de, an additional dRF pulse. Howev

ng circuits and amon makes often ner respectively. characteristics b

ments. In order to may be build d. Another way iive type of P-i-N figuration depen

ned both by that o

ver technique.

ak and average p), utilized as theve frequency reg

n for providing reTherefore, no sigever implementatgth band is complk power. Next, p

vided by such unue to a finite fohe greater such sdary emission caty short RF pulse

decrease the mendirectional couplver it is hard to

mplitude of the spnecessary utilizat

but nevertheless so increase the maby using multi

is to utilize a circ diodes should b

nds on the isolatof circulator and V

power, e input gion a eceiver gnal is tion of licated

passive nit may orward spikes. thodes e front

ntioned er and find a

pike. In tion of

special aximal i-diode culator e used tion of VSWR

Fig SuratthesumThsouswansevfolsigourecsenabnointadforIt scapspeit iswthetheproby

a) g. 12. Design of h

uch approach alloted average powee diode. The parmmarized briefly

he above design ource to its outpu

witch is inferior tod maximum rateveral switches, a llowing switches.gnal. It is importautput power of suceiver input. Evidnsitivity in this cove calibration m

ot inside the calibtroduced by temvantages of the ar wide utilizationshould be noticedpable to handle eed. It increases tis rather difficult

witches can be turere are no circuierefore the voltagocess depends st

y attitude to produ

igh power P-i-N

ows us to build aer due to the incirameters of somey in Table VII. of P-i-N switch prut terminal as depo that ended by aed power. Howev high power rati. On other hand tant if a noise souuch source is lowdently, the utilizcase. Certainly thmethod. Namely,bration loop. Seco

mperature dependabove approach a

n in the magnetrond that as followsmulti-kilowatt p

the minimal height to provide a tigrned on safely. Itits to shape the ge across magnetrongly on variouuce oscillations at

switch for receive

an absorptive swdent wave powere switches used

rovides a capabilpicted in Fig. 12,

a termination as rver, since the recing is not a mandthere are no addi

urce is used for thw enough requiration of passive

here are some dis, the input circulondly, there is sodency of P-i-N dare much more mn based radars.

s from Table VII, peak power, are ht at which the fu

ght control of thet is due to as metrailing edge of

tron drops ratherus parameters. Tht rather low anod

er protection

witch as well as tr dissipated on a in Ka band rad

lity to couple a si, b. Certainly sucrespect to both anceiver protectiondatory requiremeitional losses introhe receiver calibraring a rather highcircuits results in

sadvantages attenlator and the firsome temperature diode characterismeaningful and i

the most powerfcharacterized by

ull receiver sensite moment, when entioned above, in the output pulsr slowly (see Fig. he magnetron is gde voltages.

b)

o improve its ma termination but

dars (see Section

ignal from any exch configuration n attenuation pro

n circuitry comprent for the seconoduced for the exation. (see Fig. 10h coupling ratio n decrease in thending the utilizatst protection swit drift of the calibstics. Neverthelest can be recomm

ful switches, whiy rather low switivity is achievedthe receiver protn the most modu

se in a forced m 7) and duration generally charact

aximal not on 0) are

xternal of the ovided rises of nd and xternal 0). The to the e radar tion of tch are bration ss, the

mended

ich are itching . Next, tection ulators

manner, of this

terized

www.intechopen.com


Fig

4.3Aswhstacomproreqsimdulimbe recSois duspidetraW twThaplevconbaevicirpro

g. 11. Utilization o

3.2 Receiver pros known the maghereas maximal rage of the radar mbination of circotection. It mayquired to controlmilar approach foue to a lack of suimiters allow so-ca much less at thcovery time of thme types of magncharacterized of

uration, which mikes in the passivtector in order t

ade-off between th band high peak

wo separate antennhe switches basedproaches should

vel of both peaknfiguration. It mased switch as it iidently in the sw

rculator used, whovided by the dio

of leakage signal

otection circuitrygnetron is characrated input pow receiver, is of m

culator and limite be designed byl them. It increasor the radars operitable limiting dioalled spikes, whiche front of RF phe diode. Evidennetrons, partially

ften by a hard smay be of 2-5 nse

ve limiters provido bias the limitehe losses introdu

k power inherentnas for the transmd on P-i-N diode be used to meetk and average pakes its design rais depicted in Figwitch. Attenuati

hereas insertion loode.

to implement coh

y cterized by a rather of low noise amilliwatt level. W

er is the standard y using only pases reliability conrating within millodes accepting kch mean that the pulse than durinntly the shorter Ry the magnetronsself-excitation resec only. There arding, for example

er diode during Ruced by the biasint to the magnetromitter and receivees provide better t above requirempower a switchather complicated

g. 12, a. A reflection in such confosses is determin

herence-on-receiv

her high both peaamplifiers (LNAWithin microwav de facto solutionassive circuits. Tnsiderably. Howelimeter wavelengilowatt level pea attenuation prov

ng its flat part dRF pulse front th with cold secondsulting in a veryre methods to de, an additional dRF pulse. Howev

ng circuits and amon makes often ner respectively. characteristics b

ments. In order to may be build d. Another way iive type of P-i-N figuration depen

ned both by that o

ver technique.

ak and average p), utilized as theve frequency reg

n for providing reTherefore, no sigever implementatgth band is complk power. Next, p

vided by such unue to a finite fohe greater such sdary emission caty short RF pulse

decrease the mendirectional couplver it is hard to

mplitude of the spnecessary utilizat

but nevertheless so increase the maby using multi

is to utilize a circ diodes should b

nds on the isolatof circulator and V

power, e input gion a eceiver gnal is tion of licated

passive nit may orward spikes. thodes e front

ntioned er and find a

pike. In tion of

special aximal i-diode culator e used tion of VSWR

Fig SuratthesumThsouswansevfolsigourecsenabnointadforIt scapspeit iswthetheproby

a) g. 12. Design of h

uch approach alloted average powee diode. The parmmarized briefly

he above design ource to its outpu

witch is inferior tod maximum rateveral switches, a llowing switches.gnal. It is importautput power of suceiver input. Evidnsitivity in this cove calibration m

ot inside the calibtroduced by temvantages of the ar wide utilizationshould be noticedpable to handle eed. It increases tis rather difficult

witches can be turere are no circuierefore the voltagocess depends st

y attitude to produ

igh power P-i-N

ows us to build aer due to the incirameters of somey in Table VII. of P-i-N switch prut terminal as depo that ended by aed power. Howev high power rati. On other hand tant if a noise souuch source is lowdently, the utilizcase. Certainly thmethod. Namely,bration loop. Seco

mperature dependabove approach a

n in the magnetrond that as followsmulti-kilowatt p

the minimal height to provide a tigrned on safely. Itits to shape the ge across magnetrongly on variouuce oscillations at

switch for receive

an absorptive swdent wave powere switches used

rovides a capabilpicted in Fig. 12,

a termination as rver, since the recing is not a mandthere are no addi

urce is used for thw enough requiration of passive

here are some dis, the input circulondly, there is sodency of P-i-N dare much more mn based radars.

s from Table VII, peak power, are ht at which the fu

ght control of thet is due to as metrailing edge of

tron drops ratherus parameters. Tht rather low anod

er protection

witch as well as tr dissipated on a in Ka band rad

lity to couple a si, b. Certainly sucrespect to both anceiver protectiondatory requiremeitional losses introhe receiver calibraring a rather highcircuits results in

sadvantages attenlator and the firsome temperature diode characterismeaningful and i

the most powerfcharacterized by

ull receiver sensite moment, when entioned above, in the output pulsr slowly (see Fig. he magnetron is gde voltages.

b)

o improve its ma termination but

dars (see Section

ignal from any exch configuration n attenuation pro

n circuitry comprent for the seconoduced for the exation. (see Fig. 10h coupling ratio n decrease in thending the utilizatst protection swit drift of the calibstics. Neverthelest can be recomm

ful switches, whiy rather low switivity is achievedthe receiver protn the most modu

se in a forced m 7) and duration generally charact

aximal not on 0) are

xternal of the ovided rises of nd and xternal 0). The to the e radar tion of tch are bration ss, the

mended

ich are itching . Next, tection ulators

manner, of this

terized

www.intechopen.com


Switch #1 Switch #2 Bandwidth, MHz 300 MHz 500 MHz Maximum peak power @ 400 nsek pulse duration, W

> 3500 > 100

Insertion losses, dB < 0.7 < 0.9 Attenuation, dB > 25 > 22 VSWR in open state < 1.2 < 1.3 Switching time @ -3 dB level, nsec < 1000 < 75 Driving signal 120 mA/-25 V 50 mA/-10 V

Table VII. Parameters of switches of receiver protection circuits operating within Ka band. These oscillations are weak as usual but may cause LNA damage. A usual way to avoid that is to provide an enough delay before turning the receiver protection switches on. It leads to additional increase in the radar blind zone. The following way is utilized to minimize the response time of the receiver protection system. An additional logarithmic detector is introduced in the channel for sampling the output magnetron pulse (see Fig. 10). Due to its high sensitivity it is possible to detect extremely low level oscillations produced by the magnetron resulting in a tight control of the moment when the switches can be turned on safely. Additionally, by providing two levels detection it is possible to start turning the first – the most powerful and the slowest – switch in advance, at the moment when magnetron is still producing a high but not full power. It allows us to decrease the radar blind zone some more. It should be noticed that there is a very unpleasant effect of the magnetron operation, which affects strongly the receiver safety. Namely, some magnetrons demonstrate the attitude to start producing oscillations not during the front of the modulation pulse but during its trailing edge in opposite. In this case the delay between the triggering pulse and RF pulse may be much longer than usual and it may appear in the moment when the receiver protection circuitry is in a low attenuation state already. Thus such circuitry should process the above condition correctly. In the most of the radars in question if there is no radiation within a definite interval after the transmitter triggering pulse the protection switches remain to be open until next RF pulse.

4.3.3 Receiver dynamical range A high peak power, which are inherent to magnetron caused not only the above problem while the development of receiver protection circuitry. Another issue is related to maximization of the total receiver dynamical range. Evidently, a high peak power of the transmitter as well as high gain antennas, which can be found in high performance magnetron based radars, lead us to the necessity to introduce rather sophisticated gain regulation circuits in the receiver in order to prevent data loss under either condition. For example, the signal at the receiver input may be as strong as several milliwatts. The total dynamical range for meteorological radars similar to that described in Section 3, should exceed 110 dB. In order to prevent saturation of even the input receiver stage, a regulated attenuator should be used before it. We utilized the following approach. One of the P-i-N diodes of the receiver protection circuitry is driven in the way providing three states, namely, ‘off’ ‘on’ ‘-N dB attenuation’ respectively, instead of two states used commonly. It can be achieved by introduction of the corresponding regulation of the current through the diode or the voltage across it. Such way does not require introduction of additional

elements. Despite such approach has proven its usefulness, there were evident disadvantages. At first, the attenuation introduced depends on temperature and frequency rather strongly. Next, every time when the switch is replaced, the corresponding P-i-N diode driver circuits should be re-adjusted in order to keep a standardized attenuation level within allowable margin. An idea to use instead a step attenuator as the third switch (see Fig. 10) looks much more appreciated. However if its implementation seems to cause no serious difficulties in Ka band, for higher frequencies a dedicated development should be done to prove the value of such approach for practical utilization.

4.3.4 Automatic frequency control Since the transmitter in the radars in question utilizes a free-running magnetron, an automatic frequency control loop (AFC) should be implemented into the receiver in order to tune in it to follow the magnetron frequency. Evidently, AFC should ensure frequency locking over the whole range of possible magnetron operational frequency variation. Next, it should provide a rather good accuracy, especially if an ordinary, non-adaptive matched filtration is used during signal processing. In early radars AFC design was based on the utilization of analog circuitry including usually some implementation of a frequency discriminator. In this case it is very difficult to meet both above requirements at the same time. Actually in order to achieve a high precision measurement of frequency deviation it is needed to use resonant circuitry with a high figure of merit. On other hand it leads us to decrease in the discriminator operational bandwidth. In addition AFC should operate in a pulsed mode at rather short pulse duration. Moreover, a strong frequency deviation during leading and ending edge of the magnetron output pulse result in additional errors and their dependency on the pulse duration. Nevertheless it was possible to develop AFC on the base of an analog frequency discriminator which demonstrated acceptable performance. The following tricks were used: (i) cropping the input pulse, which reduces the contribution of the frequency modulation during leading and ending edge of the magnetron pulse to the discriminator response; (ii) a periodical throughout calibration of the discriminator circuitry with a dedicated reference signal with the same duration as for the magnetron pulse; (iii) a digital post processing of the rectified output signal of the discriminator including averaging. All above allows us to achieve the long-term AFC accuracy of ±200 kHz at the central frequency of 280 MHz and frequency locking bandwidth of ±150 MHz independently on magnetron pulse width varying within the range of 100…400 nsec. The above AFC was used in early Ka band meteorological radars (see Section 0). Continuous progresses in digital signal processing technique and direct integration of the data acquisition system in the receiver have allowed us developing a fully digital AFC circuitry. It comprises of two loops for a rough and fine frequency locking, correspondingly. The first loop is based on direct measurements of the frequency of a downconverted sampled transmitter signal (see Fig. 10) by using an ordinary digital counter. Pulse cropping is used to improve the accuracy as it was described above. This loop works over the whole range of possible magnetron frequency variations and provides the frequency locking accuracy of about ± 1 MHz at the transmitter pulse duration of 200 ns. The second loop is based on direct measurement of the phase change rate of the downconverted sampled transmitter signal, digitized by an ADC. Due to using a digital quadrature phase detector, the sign of the frequency offset can be measured, and the result of the measurements is

www.intechopen.com


Switch #1 Switch #2 Bandwidth, MHz 300 MHz 500 MHz Maximum peak power @ 400 nsek pulse duration, W

> 3500 > 100

Insertion losses, dB < 0.7 < 0.9 Attenuation, dB > 25 > 22 VSWR in open state < 1.2 < 1.3 Switching time @ -3 dB level, nsec < 1000 < 75 Driving signal 120 mA/-25 V 50 mA/-10 V

Table VII. Parameters of switches of receiver protection circuits operating within Ka band. These oscillations are weak as usual but may cause LNA damage. A usual way to avoid that is to provide an enough delay before turning the receiver protection switches on. It leads to additional increase in the radar blind zone. The following way is utilized to minimize the response time of the receiver protection system. An additional logarithmic detector is introduced in the channel for sampling the output magnetron pulse (see Fig. 10). Due to its high sensitivity it is possible to detect extremely low level oscillations produced by the magnetron resulting in a tight control of the moment when the switches can be turned on safely. Additionally, by providing two levels detection it is possible to start turning the first – the most powerful and the slowest – switch in advance, at the moment when magnetron is still producing a high but not full power. It allows us to decrease the radar blind zone some more. It should be noticed that there is a very unpleasant effect of the magnetron operation, which affects strongly the receiver safety. Namely, some magnetrons demonstrate the attitude to start producing oscillations not during the front of the modulation pulse but during its trailing edge in opposite. In this case the delay between the triggering pulse and RF pulse may be much longer than usual and it may appear in the moment when the receiver protection circuitry is in a low attenuation state already. Thus such circuitry should process the above condition correctly. In the most of the radars in question if there is no radiation within a definite interval after the transmitter triggering pulse the protection switches remain to be open until next RF pulse.

4.3.3 Receiver dynamical range A high peak power, which are inherent to magnetron caused not only the above problem while the development of receiver protection circuitry. Another issue is related to maximization of the total receiver dynamical range. Evidently, a high peak power of the transmitter as well as high gain antennas, which can be found in high performance magnetron based radars, lead us to the necessity to introduce rather sophisticated gain regulation circuits in the receiver in order to prevent data loss under either condition. For example, the signal at the receiver input may be as strong as several milliwatts. The total dynamical range for meteorological radars similar to that described in Section 3, should exceed 110 dB. In order to prevent saturation of even the input receiver stage, a regulated attenuator should be used before it. We utilized the following approach. One of the P-i-N diodes of the receiver protection circuitry is driven in the way providing three states, namely, ‘off’ ‘on’ ‘-N dB attenuation’ respectively, instead of two states used commonly. It can be achieved by introduction of the corresponding regulation of the current through the diode or the voltage across it. Such way does not require introduction of additional

elements. Despite such approach has proven its usefulness, there were evident disadvantages. At first, the attenuation introduced depends on temperature and frequency rather strongly. Next, every time when the switch is replaced, the corresponding P-i-N diode driver circuits should be re-adjusted in order to keep a standardized attenuation level within allowable margin. An idea to use instead a step attenuator as the third switch (see Fig. 10) looks much more appreciated. However if its implementation seems to cause no serious difficulties in Ka band, for higher frequencies a dedicated development should be done to prove the value of such approach for practical utilization.

4.3.4 Automatic frequency control Since the transmitter in the radars in question utilizes a free-running magnetron, an automatic frequency control loop (AFC) should be implemented into the receiver in order to tune in it to follow the magnetron frequency. Evidently, AFC should ensure frequency locking over the whole range of possible magnetron operational frequency variation. Next, it should provide a rather good accuracy, especially if an ordinary, non-adaptive matched filtration is used during signal processing. In early radars AFC design was based on the utilization of analog circuitry including usually some implementation of a frequency discriminator. In this case it is very difficult to meet both above requirements at the same time. Actually in order to achieve a high precision measurement of frequency deviation it is needed to use resonant circuitry with a high figure of merit. On other hand it leads us to decrease in the discriminator operational bandwidth. In addition AFC should operate in a pulsed mode at rather short pulse duration. Moreover, a strong frequency deviation during leading and ending edge of the magnetron output pulse result in additional errors and their dependency on the pulse duration. Nevertheless it was possible to develop AFC on the base of an analog frequency discriminator which demonstrated acceptable performance. The following tricks were used: (i) cropping the input pulse, which reduces the contribution of the frequency modulation during leading and ending edge of the magnetron pulse to the discriminator response; (ii) a periodical throughout calibration of the discriminator circuitry with a dedicated reference signal with the same duration as for the magnetron pulse; (iii) a digital post processing of the rectified output signal of the discriminator including averaging. All above allows us to achieve the long-term AFC accuracy of ±200 kHz at the central frequency of 280 MHz and frequency locking bandwidth of ±150 MHz independently on magnetron pulse width varying within the range of 100…400 nsec. The above AFC was used in early Ka band meteorological radars (see Section 0). Continuous progresses in digital signal processing technique and direct integration of the data acquisition system in the receiver have allowed us developing a fully digital AFC circuitry. It comprises of two loops for a rough and fine frequency locking, correspondingly. The first loop is based on direct measurements of the frequency of a downconverted sampled transmitter signal (see Fig. 10) by using an ordinary digital counter. Pulse cropping is used to improve the accuracy as it was described above. This loop works over the whole range of possible magnetron frequency variations and provides the frequency locking accuracy of about ± 1 MHz at the transmitter pulse duration of 200 ns. The second loop is based on direct measurement of the phase change rate of the downconverted sampled transmitter signal, digitized by an ADC. Due to using a digital quadrature phase detector, the sign of the frequency offset can be measured, and the result of the measurements is

www.intechopen.com


practically independent on the amplitude of the sampled signal. This loop ensures a much better accuracy, which can be as precise as ± 50 kHz at the output pulse duration of 200 ns, but certainly works within Nyquist range of ADC only.

4.3.5 Signal processing- a few remarks As known (see Section 0) the magnetron produces incoherent oscillations and only the introduction a corresponding signal processing procedure makes possible retrieving Doppler information. Various techniques can be used to implement such processing, see (Li et al, 1994) for example. Their discussion is not subject of matter of this paper. Nevertheless we would like to notice that a potential provided by the introduction of sophisticated signal processing are not confined by the above possibility. Instead it may change radically our mind on the principles of magnetron utilization. Let us cast a very brief look on that below. As mentioned above achievable resolution, accuracy and measurement capabilities of any magnetron based radar are related tightly to a particular implementation of sampling of the phase of radiated RF pulse. In principle such sampling can be made in one point of time only and it would be enough to implement Doppler processing in principle. But despite an evident simplicity, this method cannot be recommended due to an inappropriate quality of signal processing inherent generally to it. Instead, in our latest radars we are utilizing another approach based on the measurement of so called phase profile – i.e. set of several measurements of phase in equidistant points of time – for each transmitted pulse. Such profile is used to measure the phase shift between transmitted and received pulse by using convolution and filtration procedures. Since the phase profile is used rather than single phase measurement the following advantages are introduced: (i) a reduction of noise due to inherent averaging; (ii) the effect of intrapulse phase variation as well as the appearance of the frequency offset under downconversion are canceled and do not affect the radar sensitivity and Doppler measurement accuracy – i.e. an adaptive matched filtration is implemented in this case. The above technique can be used to implement intrapulse frequency modulation capability into magnetron based radar. Actually, it is possible to provide frequency chirp by varying the magnetron anode voltage during the modulation pulse. Then a digital technique may be utilized to calculate a correlation function for each radiated pulse and utilize a compression-like processing in order to increase the radar spatial resolution. Namely the millimeter wave magnetrons allow attaining a relatively wide frequency chirp without a significant variation of its output power. For W band magnetrons it is possible to achieve 100 MHz frequency chirp. In the case of successful implementation of the above approach, a considerable improvement both in the radar resolution and sensitivity may be expected. Next, above we described the successful implementation of precise measurement of the magnetron frequency for every radiated pulse (see Section 0). Fundamentally, the utilization of a similar technique allows us to expect a considerable improvement of Doppler processing quality in radars equipped with non-coaxial magnetrons as well as to reduce requirements for the stability of the voltage pulse across the magnetron. Actually, the usage of a high resolution ADC increases greatly the precision of frequency measurement. Certainly we mean relative measurements. We expect that for 200 ns pulse length the achievable accuracy may be as lower as 1 kHz that corresponds to relative accuracy of 3·10-8 and 10-8 for Ka and W band magnetrons respectively. The indicated value is on about order better than the best frequency stability achieved (see Section 0) resulting in the

corresponding improvement of Doppler processing quality. Certainly sophisticated algorithm should be developed to avoid influence of noise, intrapulse frequency modulation etc on the frequency measurements accuracy. This approach is important especially for W band magnetrons, which frequency stability is much less than for Ka band devices and coaxial devices are not produced at all.

5. Conclusion

The above discussion demonstrates clearly that the utilization of recent advances in magnetron manufacturing technology, the introduction of novel approaches in radar design as well as a vast progress in digital signal processing technique result in a solid overall performance of the magnetron based millimeter wavelengths radars. A number of such radars developed and manufactured by our institution, which have been put into operation during last several years as well as the experience of their utilization allow us to claim for a great potential of the magnetrons for the radar development. We dare to affirm that a rather inveterate opinion about futility of such devices to built contemporary radar systems should be considered to be expired. An evident room for ulterior improvements in both developing high voltage modulators ensuring tighter control of the magnetron operation and introducing sophisticated signal processing algorithms make possible to expect even higher performance level of the magnetron based radars, especially for that operating within short millimeter wavelength bands in the nearest future.

6. Acknowledgments

I am appreciating vastly to Prof. D. M. Vavriv and Dr. V. D. Naumenko for a fruitful discussions and comprehensive assistance.

7. References

Barker et al. (2005). Modern Microwave and Millimeter-Wave Power Electronics. Wiley-IEEE Press. Belikov et all. (2002). 95 GHz, 2 kW magnetron transmitter of 20 ns pulses. Proceedings of

German Radar Symposium-2002, (стр. 571-574). Bonn. Brown. (1999). Technical and Military Imperatives: A Radar History of World War 2. New York:

Taylor & Francis. Gritsaenko, S.; Eremka, V.; Kopot, M.; Kulagin, O.; Naumenko, V.; Suvorov, A. (2005). Multi-

cavity magnetrons with cold secondary emission cathode: achievments, problems, prospects (in Russian). Radio Physics & Electronics , 10 (Secial issue), 499-529.

Jenett, M.; Kazantsev, V.; Kurekin, A.; Schunemann, K.; Vavriv, D.; Vinogradov, V.; Volkov, V. (1999). Dual 94 and 36 GHz radar system for remote sensing applications. Proceedings of International Geoscience and Remote Sensing Symposium, 5, стр. 2596-2598. Hamburg.

Li, Hua; Illingworth, A. J.; Eastment, J (1994) A simple method of Dopplerizing a pulsed magnetron radar. Microwave Journal , 37 (4), 226-235.

Okress. (1961). Crossed-field Microwave Devices (Т. 1,2). New York & London: Academic Press. Sivan. (1994). Microwave Tube Transmitters. London: Chapman&Hill.

www.intechopen.com


practically independent on the amplitude of the sampled signal. This loop ensures a much better accuracy, which can be as precise as ± 50 kHz at the output pulse duration of 200 ns, but certainly works within Nyquist range of ADC only.

4.3.5 Signal processing- a few remarks As known (see Section 0) the magnetron produces incoherent oscillations and only the introduction a corresponding signal processing procedure makes possible retrieving Doppler information. Various techniques can be used to implement such processing, see (Li et al, 1994) for example. Their discussion is not subject of matter of this paper. Nevertheless we would like to notice that a potential provided by the introduction of sophisticated signal processing are not confined by the above possibility. Instead it may change radically our mind on the principles of magnetron utilization. Let us cast a very brief look on that below. As mentioned above achievable resolution, accuracy and measurement capabilities of any magnetron based radar are related tightly to a particular implementation of sampling of the phase of radiated RF pulse. In principle such sampling can be made in one point of time only and it would be enough to implement Doppler processing in principle. But despite an evident simplicity, this method cannot be recommended due to an inappropriate quality of signal processing inherent generally to it. Instead, in our latest radars we are utilizing another approach based on the measurement of so called phase profile – i.e. set of several measurements of phase in equidistant points of time – for each transmitted pulse. Such profile is used to measure the phase shift between transmitted and received pulse by using convolution and filtration procedures. Since the phase profile is used rather than single phase measurement the following advantages are introduced: (i) a reduction of noise due to inherent averaging; (ii) the effect of intrapulse phase variation as well as the appearance of the frequency offset under downconversion are canceled and do not affect the radar sensitivity and Doppler measurement accuracy – i.e. an adaptive matched filtration is implemented in this case. The above technique can be used to implement intrapulse frequency modulation capability into magnetron based radar. Actually, it is possible to provide frequency chirp by varying the magnetron anode voltage during the modulation pulse. Then a digital technique may be utilized to calculate a correlation function for each radiated pulse and utilize a compression-like processing in order to increase the radar spatial resolution. Namely the millimeter wave magnetrons allow attaining a relatively wide frequency chirp without a significant variation of its output power. For W band magnetrons it is possible to achieve 100 MHz frequency chirp. In the case of successful implementation of the above approach, a considerable improvement both in the radar resolution and sensitivity may be expected. Next, above we described the successful implementation of precise measurement of the magnetron frequency for every radiated pulse (see Section 0). Fundamentally, the utilization of a similar technique allows us to expect a considerable improvement of Doppler processing quality in radars equipped with non-coaxial magnetrons as well as to reduce requirements for the stability of the voltage pulse across the magnetron. Actually, the usage of a high resolution ADC increases greatly the precision of frequency measurement. Certainly we mean relative measurements. We expect that for 200 ns pulse length the achievable accuracy may be as lower as 1 kHz that corresponds to relative accuracy of 3·10-8 and 10-8 for Ka and W band magnetrons respectively. The indicated value is on about order better than the best frequency stability achieved (see Section 0) resulting in the

corresponding improvement of Doppler processing quality. Certainly sophisticated algorithm should be developed to avoid influence of noise, intrapulse frequency modulation etc on the frequency measurements accuracy. This approach is important especially for W band magnetrons, which frequency stability is much less than for Ka band devices and coaxial devices are not produced at all.

5. Conclusion

The above discussion demonstrates clearly that the utilization of recent advances in magnetron manufacturing technology, the introduction of novel approaches in radar design as well as a vast progress in digital signal processing technique result in a solid overall performance of the magnetron based millimeter wavelengths radars. A number of such radars developed and manufactured by our institution, which have been put into operation during last several years as well as the experience of their utilization allow us to claim for a great potential of the magnetrons for the radar development. We dare to affirm that a rather inveterate opinion about futility of such devices to built contemporary radar systems should be considered to be expired. An evident room for ulterior improvements in both developing high voltage modulators ensuring tighter control of the magnetron operation and introducing sophisticated signal processing algorithms make possible to expect even higher performance level of the magnetron based radars, especially for that operating within short millimeter wavelength bands in the nearest future.

6. Acknowledgments

I am appreciating vastly to Prof. D. M. Vavriv and Dr. V. D. Naumenko for a fruitful discussions and comprehensive assistance.

7. References

Barker et al. (2005). Modern Microwave and Millimeter-Wave Power Electronics. Wiley-IEEE Press. Belikov et all. (2002). 95 GHz, 2 kW magnetron transmitter of 20 ns pulses. Proceedings of

German Radar Symposium-2002, (стр. 571-574). Bonn. Brown. (1999). Technical and Military Imperatives: A Radar History of World War 2. New York:

Taylor & Francis. Gritsaenko, S.; Eremka, V.; Kopot, M.; Kulagin, O.; Naumenko, V.; Suvorov, A. (2005). Multi-

cavity magnetrons with cold secondary emission cathode: achievments, problems, prospects (in Russian). Radio Physics & Electronics , 10 (Secial issue), 499-529.

Jenett, M.; Kazantsev, V.; Kurekin, A.; Schunemann, K.; Vavriv, D.; Vinogradov, V.; Volkov, V. (1999). Dual 94 and 36 GHz radar system for remote sensing applications. Proceedings of International Geoscience and Remote Sensing Symposium, 5, стр. 2596-2598. Hamburg.

Li, Hua; Illingworth, A. J.; Eastment, J (1994) A simple method of Dopplerizing a pulsed magnetron radar. Microwave Journal , 37 (4), 226-235.

Okress. (1961). Crossed-field Microwave Devices (Т. 1,2). New York & London: Academic Press. Sivan. (1994). Microwave Tube Transmitters. London: Chapman&Hill.

www.intechopen.com


Skolnik, (2008). Radar Handbook, Third Edition. New York: McGraw-Hill. Tsimring. (2007). Electron Beams and Microwave Vacuum Electronics. Hoboken, NJ:

John Wiley & Sons, Inc. Usikov. (1972). Investigationgs in Microwave Electronics made in IRE AS Ukraine (in

Russian). Electronic Technics, Ser 1, #12 (12), 39-49. Vavriv et all. (2002). 95 GHz Doppler polarimetric cloud radar based on a magnetron transmitter.

Proceedings of 32nd Microwave European Conference, 1, стр. 665 -668. Milan. Volkov et all. (2007). A Ka-band, Magnetron based, scanning radar for airborne applications.

Radar Conference, 2007. EuRAD 2007, (стр. 186-189). Munich. Zhu, Di; Dong, Xiaolong; Lin, Wenming (2008). Pulse Compression with Very Low

Sidelobes in a Spaceborne Weather Radar, Proceedings of Geosciences and Remote Sensing Symposium, IGARSS 2008, Vol. 5, pp. 252-255, ISBN: 978-1-4244-2807-6, Jul. 2008, Boston, USA

www.intechopen.com

Microwave and Millimeter Wave Technologies Modern UWBantennas and equipmentEdited by Igor Mini

ISBN 978-953-7619-67-1Hard cover, 488 pagesPublisher InTechPublished online 01, March, 2010Published in print edition March, 2010

InTech EuropeUniversity Campus STeP Ri Slavka Krautzeka 83/A 51000 Rijeka, Croatia Phone: +385 (51) 770 447 Fax: +385 (51) 686 166www.intechopen.com

InTech ChinaUnit 405, Office Block, Hotel Equatorial Shanghai No.65, Yan An Road (West), Shanghai, 200040, China

Phone: +86-21-62489820 Fax: +86-21-62489821

How to referenceIn order to correctly reference this scholarly work, feel free to copy and paste the following:

Vadym Volkov (2010). Magnetron Based Radar Systems for Millimeter Wavelength Band – ModernApproaches and Prospects, Microwave and Millimeter Wave Technologies Modern UWB antennas andequipment, Igor Mini (Ed.), ISBN: 978-953-7619-67-1, InTech, Available from:http://www.intechopen.com/books/microwave-and-millimeter-wave-technologies-modern-uwb-antennas-and-equipment/magnetron-based-radar-systems-for-millimeter-wavelength-band-modern-approaches-and-prospects

Documents

Vadym Volkov Magnetron Based Radar Systems for …cdn.intechweb.org/pdfs/10315.pdf · Magnetron Based Radar Systems for Millimeter Wavelength Band – Modern Approaches and Prospects